System and method for virtual inspection and response planning for wildfire ignition risk reduction

Abstract

A computer-implemented method for tracking wildfire mitigations. A three-dimensional (3D) reconstruction of the outside walls of a building structure and tagged objects are in this 3D reconstruction are simultaneously displayed on a display screen of a mobile computing device. In response to a selection of a tagged object, the display screen directs the service provider to a stored physical location associated with the tagged object while geospatially tracking the mobile computing device. The service provider can be prompted to confirm that the work is complete. Systems for performing a virtual inspection of a building structure are also provided. One example system includes a data store and a virtual inspection engine. Another example system includes a first data store, an image processing engine, a second data store, a wildfire risk modeling engine, an annotation engine, and a post-processing engine. A virtual wildfire inspection apparatus is also provided.

Description

Attorney Docket No. FWIG-003W001 SYSTEM AND METHOD FOR VIRTUAL INSPECTION AND RESPONSE PLANNING FOR WILDFIRE IGNITION RISK REDUCTIONCROSS-REFERENCE TO RELATED APPLICATION(S)

[0001] This application claims priority to U.S. Provisional Application No. 63 / 744,028, titled “Method For Structure Response Planning For Wildfire Ignition Risk Reduction,” filed on January 10, 2025, which is hereby incorporated herein by reference.BACKGROUND

[0002] Wildfires are an increasingly important factor for homeowners to consider in various parts of the world, including in North America. In high-risk areas, wildfires present a real danger of total loss of a building structure, such as a residential house. Assessing a building structure and corresponding land for wildfire requires that a skilled technician travel to the property to take measurements of the building structure including those of structural features on the building structure, such as windows, doors, vents, soffits, and decks. The technician also needs to evaluate the condition of the structural features, for example to determine whether wildfire mitigation measures have been implemented to minimize wildfire risks. Current at-location wildfire mitigation plans are lists of general or site-specific tasks to be completed in preparation for an impending wildfire. In some cases mitigations are determined on the spot by a responder. The current approach has significant drawbacks in that the response-planning personnel and the service personnel are almost always different causing a large information gap in translating checklist actions into instructions. Any response in preparation to a wildfire is often extremely time constrained and having an inconsistent on incomplete response can result in structure loss.SUMMARY

[0003] In some aspects, the techniques described herein relate to a computer-implemented method for tracking wildfire mitigation actions, including: displaying, on a display screen of a mobile computing device associated with a user, a displayed portion of a volumetrically rendered three-dimensional reconstruction of outside walls of a building structure; displaying, on the display screen, one or more tagged objects, each tagged object representing a respective physical object and / or a respective physical area where one or more assigned wildfire mitigation actions is / are to be performed; receiving, with the mobile computingAttorney Docket No. FWIG-003W001 device, a selection of one of the one or more tagged objects to form a selected tagged object; providing, with the mobile computing device, a physical location associated with the selected tagged object; providing, with the mobile computing device, the one or more assigned wildfire mitigation actions associated with the selected tagged object; geospatially tracking the mobile computing device while the user performs the one or more assigned wildfire mitigation actions associated with the selected tagged object; requesting, with the mobile computing device, a confirmation that the user performed the one or more assigned wildfire mitigation actions associated with the selected tagged object; receiving, with the mobile computing device, the confirmation that the user performed the one or more assigned wildfire mitigation actions associated with the selected tagged object; updating, in response to the confirmation and with the mobile computing device, a status of the selected tagged object; and requesting, in response to the confirmation and with the mobile computing device, that the user capture digital images of all the outside walls of the building structure including the physical location associated with the selected tagged object.

[0004] In some aspects, the techniques described herein relate to a method, further including providing, with the mobile computing device, a geolocation associated with the selected tagged object.

[0005] In some aspects, the techniques described herein relate to a method, further including: determining a current geolocation of the mobile computing device; and providing directions, with the mobile computing device, to direct the user from the current geolocation to the geolocation associated with the selected tagged object.

[0006] In some aspects, the techniques described herein relate to a method, further including receiving, with the mobile computing device, one or more user inputs representing the confirmation that the user performed the one or more assigned wildfire mitigation actions associated with the selected tagged object.

[0007] In some aspects, the techniques described herein relate to a method, wherein the step of requesting, with the mobile computing device, the confirmation that the user performed the one or more assigned wildfire mitigation actions associated with the selected tagged object includes prompting the user, with the mobile computing device, to capture one or more digital images of the respective physical object and / or the respective physical area associated with the selected tagged object, the one or more digital images representing a completion ofAttorney Docket No. FWIG-003W001 the one or more assigned wildfire mitigation actions associated with the selected tagged object.

[0008] In some aspects, the techniques described herein relate to a method, further including: capturing the one or more digital images of the respective physical object and / or the respective physical area associated with the selected tagged object, the one or more digital images representing the completion of the one or more assigned wildfire mitigation actions associated with the selected tagged object; and sending the one or more digital images from the mobile computing device to a second computer in communication with the mobile computing device.

[0009] In some aspects, the techniques described herein relate to a method, wherein the step of updating, in response to the confirmation and with the mobile computing device, the status of the selected tagged object includes changing a visual appearance of the selected tagged object on the display screen.

[0010] In some aspects, the techniques described herein relate to a system for performing a virtual inspection of a building structure on a property to determine a risk of loss due to wildfire, the building structure having outside walls, the system including: a data store that includes image data that represent a geospatially registered three-dimensional rendering of the outside walls of the building structure and property features data that represent tagged objects, the property features data including indexed image positions and corresponding indexed geospatial positions of the tagged objects; a virtual inspection engine in communication with the data store, the virtual inspection engine configured to: display a displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the building structure according to a position and field-of-view (FOV) of a virtual camera relative to the geospatially registered three-dimensional rendering of the outside walls of the building structure, the displayed portion including one or more of the tagged objects, overlay one or more two-dimensional bounding regions over the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the building structure, each of the one or more two-dimensional bounding regions having a respective indexed image position corresponding to a respective one or more of the tagged objects, the respective indexed image position included in the property features data, receive a user input that represents a modification to one or more properties of a target two-dimensional bounding region for a target tagged object, and send an update to the data store to update the propertyAttorney Docket No. FWIG-003W001 features data according to the modification to the one or more properties of the target two-dimensional bounding region.

[0011] In some aspects, the techniques described herein relate to a system, wherein the modification to the one or more properties of the target two-dimensional bounding region includes a position, a size, and / or a shape of the target two-dimensional bounding region.

[0012] In some aspects, the techniques described herein relate to a system, wherein the update to the property features data includes an update to the respective indexed image position of the target tagged object and a corresponding update to an indexed geospatial position of the target tagged object.

[0013] In some aspects, the techniques described herein relate to a system, wherein the update to the property features data includes an update to a respective indexed image size of the target tagged object and a corresponding update to an indexed geospatial size of the target tagged object, the indexed geospatial size representing a virtual measurement of dimensions of the target tagged object.

[0014] In some aspects, the techniques described herein relate to a system wherein: the user input is a first user input, and the virtual inspection engine is further configured to: receive a second user input that represents a tag of an untagged object represented in the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the building structure, the tag of the untagged object resulting in a new tagged object, define a new two-dimensional bounding region for the new tagged object, the new two-dimensional bounding region having a new indexed image position, overlay the new two-dimensional bounding region over the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the building structure, wherein the update to the data store includes an addition of the new tagged object to the property features data.

[0015] In some aspects, the techniques described herein relate to a system, wherein: the virtual inspection engine is further configured to receive annotation data for the new tagged object, and the update to the data store includes the annotation data for the new tagged object.

[0016] In some aspects, the techniques described herein relate to a system, wherein the system further includes an image processor having an image recognition engine configured to: automatically detect structural features of the building structure in the geospatially registered three-dimensional rendering of the outside walls of the building structure, and map the structural features automatically detected by the image recognition engine to an imageAttorney Docket No. FWIG-003W001 space of the geospatially registered three-dimensional rendering of the outside walls of the building structure, wherein the image processor is configured to store the structural features automatically detected by the image recognition engine as at least some of the tagged objects in the property features data stored in the data store.

[0017] In some aspects, the techniques described herein relate to a system, wherein the virtual inspection engine includes: a rendering engine configured to produce output data that represent the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the building structure, and a tagged object rendering engine configured to generate each of the one or more two-dimensional bounding regions according to the respective indexed image position corresponding to the respective one or more of the tagged objects.

[0018] In some aspects, the techniques described herein relate to a system, wherein: the user input is a first user input, the virtual inspection engine further includes a user input detector configured to detect a second user input that represents a change in the position and / or the FOV of the virtual camera, the rendering engine is configured to update the output data that represent the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the building structure according to the change in the position and / or the FOV of the virtual camera, and the tagged object rendering engine is configured to update the one or more two-dimensional bounding regions generated according to the change in the position and / or the FOV of the virtual camera.

[0019] In some aspects, the techniques described herein relate to a system, wherein: the image data in the data store further represents an overhead rendering of a target property that includes the building structure, the one or more of the tagged objects displayed in the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the building structure is / are one or more first tagged objects, the one or more bounding regions is / are one or more first bounding regions, the virtual inspection engine is further configured to overlay the overhead rendering of the target property over the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the building structure, the overhead rendering including one or more second bounding regions for one or more second tagged objects represented in the overhead rendering.Attorney Docket No. FWIG-003W001

[0020] In some aspects, the techniques described herein relate to a system, wherein: the one or more first tagged objects is / are one or more structural features of the building structure, and the one or more second tagged objects is / are one or more trees in the target property.

[0021] In some aspects, the techniques described herein relate to a virtual wildfire inspection apparatus including: a display screen; one or more microprocessors in communication with the display screen; non-transitory computer memory in communication with the one or more microprocessors, the non-transitory computer memory storing property features data for a target building structure on a target property, the property features data including: image data representing a volumetrically rendered geospatially calibrated 3D reconstruction of the outside walls of the target building structure, and tagged object data representing tagged objects on the outside walls of the target building structure, the tagged object data including a respective indexed position of each tagged object and annotation data for each tagged object, the non-transitory computer memory further storing computer-readable instructions that, when executed by the one or more microprocessors, cause the one or more microprocessors to run a virtual inspection engine configured to: display, using the image data, a displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the target building structure according to a position and field-of-view (FOV) of a virtual camera relative to the geospatially registered three-dimensional rendering of the outside walls of the target building structure, the displayed portion including one or more of the tagged objects, overlay one or more two-dimensional bounding regions over the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the target building structure, each of the one or more two-dimensional bounding regions having a respective indexed image position corresponding to a respective one or more of the tagged objects; receive a user input that represents a modification to one or more properties of a target two-dimensional bounding region for a target tagged object, and generate an update to the property features data stored in the non-transitory computer memory according to the modification to the one or more properties of the target two-dimensional bounding region.

[0022] In some aspects, the techniques described herein relate to an apparatus, wherein the modification to the one or more properties of the target two-dimensional bounding region includes a position, a size, and / or a shape of the target two-dimensional bounding region.

[0023] In some aspects, the techniques described herein relate to an apparatus, wherein the update to the property features data includes an update to the respective indexed imageAttorney Docket No. FWIG-003W001 position of the target tagged object and a corresponding update to an indexed geospatial position of the target tagged object.

[0024] In some aspects, the techniques described herein relate to an apparatus, wherein the update to the property features data includes an update to a respective indexed image size of the target tagged object and a corresponding update to an indexed geospatial size of the target tagged object, the indexed geospatial size representing a virtual measurement of dimensions of the target tagged object.

[0025] In some aspects, the techniques described herein relate to an apparatus, wherein: the user input is a first user input, and the virtual inspection engine is further configured to: receive a second user input that represents a tag of an untagged object represented in the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the target building structure, the tag of the untagged object resulting in a new tagged object, define a new two-dimensional bounding region for the new tagged object, the new two-dimensional bounding region having a new indexed image position, overlay the new two-dimensional bounding region over the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the target building structure, wherein the update to the property features data includes an addition of the new tagged object to the property features data.

[0026] In some aspects, the techniques described herein relate to an apparatus, wherein: the virtual inspection engine is further configured to receive annotation data for the new tagged object, and the update to the property features data includes the annotation data for the new tagged object.

[0027] In some aspects, the techniques described herein relate to an apparatus, wherein: the image data further represents an overhead rendering of the target property, the one or more of the tagged objects displayed in the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the target building structure is / are one or more first tagged objects, the one or more bounding regions is / are one or more first bounding regions, the virtual inspection engine is further configured to overlay the overhead rendering of the target property and the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the target building structure, the overhead rendering including one or more second bounding regions for one or more second tagged objects represented in the overhead rendering.Attorney Docket No. FWIG-003W001

[0028] In some aspects, the techniques described herein relate to a system for determining a risk of loss to a building structure on a property due to wildfire, the building structure having outside walls, the system including: a first data store that receives and stores image data from one or more image sources, the image data including a collective representation of at least a portion of the property and all of the outside walls of the building structure including structural features on and / or in the outside walls and a geospatially calibrated overhead view of the property including the building structure; an image processing engine in communication with the first data store, the image processing engine including one or more image processors, the image processing engine configured to: detect the outside walls of the building structure and the structural features on and / or in the outside walls; generate encoded structural descriptors that include one or more properties of the outside walls of the building structure, one or more properties of the structural features on and / or in the outside walls, and indexed structural locations of the outside walls and the structural features; detect one or more fuel sources on the at least a portion of the property; generate encoded fuel descriptors that include one or more properties of each of the one or more fuel sources and one or more indexed fuel locations of the one or more fuel sources, respectively; a second data store that receives and stores the encoded structural descriptors and the encoded fuel descriptors, the second data store in communication with the image processing engine; a wildfire risk modeling engine in communication with the second data store, the wildfire risk modeling engine including one or more risk processors configured to execute computer-readable instructions corresponding to a wildfire loss risk model, the wildfire risk modeling engine receiving the encoded structural descriptors and the encoded fuel descriptors as inputs and producing at least one output; an annotation engine in communication with the wildfire risk modeling engine, the first data store, and the second data store, the annotation engine including one or more annotation processors, the annotation engine configured to: display, using the image data, a geospatially registered rendering of the outside walls of the building structure and objects on the at least a portion of the property, the objects including uncategorized objects and the one or more fuel sources; generate, using the encoded structural descriptors, structural bounding regions, each structural bounding region corresponding to a respective indexed structural location of a respective structural feature; generate, using the encoded fuel source descriptors, one or more fuel source bounding regions, each fuel source bounding region corresponding to a respective indexed fuel location of a respective fuel source; overlay the structural bounding regions and the one or more fuel source bounding regions on the geospatially registered rendering, the structural boundingAttorney Docket No. FWIG-003W001 regions and the one or more fuel source bounding regions configured to be adjustable in response to user input; send an update to the second data store to modify at least one of the indexed structural locations and / or the one or more indexed fuel locations based on the user input, wherein the wildfire risk modeling engine is configured to update the at least one output in response to the update to the second data store.

[0029] In some aspects, the techniques described herein relate to a system, wherein the geospatially registered rendering of the outside walls of the building structure includes a geospatially registered three-dimensional rendering of the outside walls of the building structure.

[0030] In some aspects, the techniques described herein relate to a system, wherein the geospatially registered rendering of the outside walls of the building structure includes an overhead rendering of the property.

[0031] In some aspects, the techniques described herein relate to a system, wherein: the user input is a first user input, the annotation engine is further configured to receive annotation data, in response to a second user input, the annotation data that include annotations that update and / or supplement the one or more properties of the structural features on and / or in the outside walls of the building structure, wherein the update to the second data store includes the annotation data.

[0032] In some aspects, the techniques described herein relate to a system, wherein: the user input is a first user input, the annotation engine is further configured to: receive a second user input that defines a new structural bounding region and annotated structural descriptors for a first uncategorized object of the uncategorized objects, determine a new indexed structural location for the uncategorized object, the first indexed structural location corresponding to the new structural bounding region, and send an update to the second data store to add the new indexed structural location to the encoded structural descriptors to produce updated structural descriptors, wherein the wildfire risk modeling engine is configured to update the at least one output based on the updated encoded structural descriptors.

[0033] In some aspects, the techniques described herein relate to a computer-implemented method for managing property assessment data for wildfire risk assessment, the method including: A) storing, in a non-transitory computer memory, a property features data structure representing a property including one or more building structures, the property features data structure including: structures data containing one or more structure objects, each structureAttorney Docket No. FWIG-003W001 object representing a respective building structure of the one or more building structures and including a structure shape defined as a polygon in geospatial coordinates, a structure type field indicating whether the respective building structure is a main structure or a fuel structure, and walls data containing one or more wall objects; trees data containing one or more tree objects, each tree object representing a respective tree on or near the property and including a tree shape defined as a polygon representing a canopy footprint of the respective tree, a height value, and trim information; and an elevation section containing terrain elevation data including geographic bounds and a plurality of elevation values representing elevation above sea level; B) receiving, from a virtual inspection application executed by one or more processors, a modification to one or more values in the property features data structure, the modification representing a correction identified by an operator; C) generating, with the one or more processors, a differential change record representing the modification, the differential change record including: a path specifying a hierarchical location within the property features data structure where the modification is to be applied, an operation type selected from add, replace, or remove, and a new value for the modification when the operation type is add or replace; D) storing the differential change record in the non-transitory computer memory separately from the property features data structure; and E) applying the differential change record to the property features data structure to produce an updated property features data structure.

[0034] In some aspects, the techniques described herein relate to a method, wherein the property features data structure further includes an info section containing property metadata, the property metadata including: an address of the property, a center of analysis represented as latitude and longitude coordinates, and risk data aggregated from a plurality of authoritative sources.

[0035] In some aspects, the techniques described herein relate to a method, wherein the risk data includes a plurality of risk values from two or more sources selected from: a public utilities commission, a state fire protection agency, and a federal agriculture agency, and wherein each risk value of the plurality of risk values includes: a numeric risk value, a risk label, a source identifier identifying a source of the risk value, and a load version indicating a date or time when the risk value was obtained.

[0036] In some aspects, the techniques described herein relate to a method, wherein the property features data structure is stored as a JSON file using a WGS84 / EPSG:4326 coordinate reference system for the geospatial coordinates.Attorney Docket No. FWIG-003W001

[0037] In some aspects, the techniques described herein relate to a method, further including: F) tracking a workflow state of the property features data structure using a state machine, the state machine defining a plurality of states including a ready -for-repair state, a processing state, and a ready-for-review state; G) transitioning the workflow state from the ready-for-repair state to the processing state in response to receiving the differential change record; and H) transitioning the workflow state from the processing state to the ready-for-review state in response to successfully applying the differential change record to produce the updated property features data structure.

[0038] In some aspects, the techniques described herein relate to a method, wherein when applying the differential change record fails due to an intermediate change to the property features data structure or a schema incompatibility, the workflow state reverts to the ready -for-repair state and an error notification is generated.

[0039] In some aspects, the techniques described herein relate to a method, further including: F) receiving a plurality of differential change records; G) detecting a conflict when two differential change records of the plurality of differential change records target a same element within the property features data structure with logically inconsistent values; and H) in response to detecting the conflict, generating an error report and rejecting the two differential change records.

[0040] In some aspects, the techniques described herein relate to a method, wherein: the property features data structure further includes other shapes data containing one or more other shape objects, each other shape object representing a miscellaneous property feature not formally categorized in the structures data or the trees data, and wherein the miscellaneous property feature is a deck, a fence, a fuel source, a fire hydrant, a site access route, a spray plan, or a retardant container location, and the method further includes: F) creating a modified copy of the property features data structure, the modified copy reflecting an expected state of the property after one or more proposed mitigation actions; G) executing a property ignition model on the property features data structure to calculate a first probability of destruction; H) executing the property ignition model on the modified copy to calculate a second probability of destruction; I) computing a risk reduction value as a difference between the first probability of destruction and the second probability of destruction, the risk reduction value quantifying a benefit of the one or more proposed mitigation actions; J) creating a plurality of modified copies of the property features data structure, each modified copy of the plurality of modified copies reflecting a different set of proposed mitigation actions; K)Attorney Docket No. FWIG-003W001 executing the property ignition model on each modified copy of the plurality of modified copies to calculate a respective probability of destruction for each modified copy; and L) generating a comparison report showing a respective risk reduction value for each modified copy, enabling selection of proposed mitigation actions based on risk reduction benefit.

[0041] In some aspects, the techniques described herein relate to a computer-implemented method for representing structural features along a wall of a building structure, the method including: A) storing, in a non-transitory computer memory, a wall object representing a wall segment of the building structure, the wall object including a wall linestring defining the wall segment in geospatial coordinates, the wall linestring having a start point and an end point; B) storing, in the non-transitory computer memory and in association with the wall object, one or more wall span objects, each wall span object of the one or more wall span objects representing a respective structural feature along the wall segment and including: a span type indicating a type of the respective structural feature, a percentage start value representing a starting position of the respective structural feature as a percentage of a length of the wall linestring, and a percentage end value representing an ending position of the respective structural feature as a percentage of the length of the wall linestring; C) receiving, at one or more processors, a modification to the wall linestring that changes at least one of the start point or the end point of the wall linestring; and D) in response to the modification to the wall linestring, automatically maintaining relative positions of the respective structural features along the wall segment by: preserving the percentage start value and the percentage end value for each wall span object of the one or more wall span objects, and computing updated geospatial coordinates for each respective structural feature by interpolating along a modified wall linestring according to the preserved percentage start value and the preserved percentage end value for each wall span object.

[0042] In some aspects, the techniques described herein relate to a method, wherein each wall span object further includes one or more of: a material designation indicating a material of the respective structural feature, a glass type indicating a type of glass for the respective structural feature when the span type is window, a roof height value indicating a height of a roof at the wall segment, one or more image references associating the respective structural feature with one or more images, and one or more observation references associating the respective structural feature with one or more inspection observations.Attorney Docket No. FWIG-003W001

[0043] In some aspects, the techniques described herein relate to a method, wherein the percentage start value and the percentage end value are each constrained to values in a range from 0 to 100, inclusive.

[0044] In some aspects, the techniques described herein relate to a method, wherein the wall object is stored in walls data of a structure object, the structure object representing the building structure and being stored in structures data of a property features data structure, and wherein the property features data structure further includes trees data and an elevation section.

[0045] In some aspects, the techniques described herein relate to a system for managing property assessment data for wildfire risk calculation, the system including: a data store configured to store a property features data structure, the property features data structure including: structure data representing one or more building structures on a property, the structure data including, for each building structure of the one or more building structures, a structure shape defined as a polygon in geospatial coordinates, a structure type field indicating whether the building structure is a main structure or a fuel structure, wall data representing one or more wall segments of the building structure, and roof data representing a roof of the building structure; tree data representing one or more trees on or near the property, the tree data including, for each tree of the one or more trees, a tree shape defined as a polygon representing a canopy footprint, a height value, a canopy radius value, and trim information; and elevation data representing terrain elevation across the property and including geographic bounds and an plurality of elevation values; a landscape pipeline implemented by one or more first processors in communication with the data store, the landscape pipeline configured to: receive input data including an address of the property, satellite imagery of the property, and parcel data, generate an initial property features data structure using the input data, and store the initial property features data structure in the data store; a property application implemented by one or more second processors in communication with the data store, the property application configured to: display a rendering of the property based on the property features data structure, receive user input representing one or more corrections to the property features data structure, and generate one or more differential change records representing the one or more corrections; and a property ignition model implemented by one or more third processors in communication with the data store, the property ignition model configured to: receive the property features data structure as input, and calculate a probability of destruction of at least one building structure of the oneAttorney Docket No. FWIG-003W001 or more building structures due to wildfire based at least in part on the structure data, the tree data, and the elevation data.

[0046] In some aspects, the techniques described herein relate to a system, wherein the trim information for each tree of the one or more trees includes: a maximum radius reduction value representing a maximum amount by which the canopy radius value can be reduced by trimming, an actual trim height value representing a current trim height of the tree based on observed cut scars, a maximum trim height value representing a maximum height to which the tree can be trimmed, and a highest adjacent roofline value representing a height of a highest roofline of structures adjacent to the tree.

[0047] In some aspects, the techniques described herein relate to a system, wherein each tree of the one or more trees further includes a breast height diameter value representing a diameter of a trunk of the tree measured at a standard forestry measurement height.

[0048] In some aspects, the techniques described herein relate to a system, wherein the property ignition model is further configured to use the breast height diameter value to calculate an estimated heat production of the tree during combustion based on a correlation between trunk diameter and thermal energy output.

[0049] In some aspects, the techniques described herein relate to a system, wherein the property ignition model is further configured to weight a tree of the one or more trees as a lower ignition risk when the tree is positioned at a higher elevation than the at least one building structure based on the elevation data.

[0050] In some aspects, the techniques described herein relate to a system, wherein the tree data for each tree of the one or more trees further includes: a natural crown radius value representing an untrimmed canopy radius, a natural lowest branch height value representing a height of a lowest branch in an untrimmed state, and a species designation selected from a controlled vocabulary based on tree shape characteristics.

[0051] In some aspects, the techniques described herein relate to a system, wherein the structure type field is set to one of: main, indicating the building structure is a primary structure to be protected from wildfire, or fuel, indicating the building structure is a neighboring structure or outbuilding that may contribute to wildfire ignition risk.

[0052] In some aspects, the techniques described herein relate to a system, wherein the roof data for each building structure includes a roof object, the roof object including: a roof shapeAttorney Docket No. FWIG-003W001 defined as a polygon in geospatial coordinates, materials data containing one or more roofing material classifications, and features data containing one or more roof feature objects.

[0053] In some aspects, the techniques described herein relate to a system, wherein each roof feature object of the one or more roof feature objects includes a feature type field set to one of chimney, skylight, or vent, and wherein when the feature type field is set to vent, the roof feature object further includes a vent type field and a vent subtype field.

[0054] In some aspects, the techniques described herein relate to a system, wherein the wall data for each building structure includes one or more wall objects, each wall object including: a wall linestring defining a wall segment in geospatial coordinates, the wall linestring having a start point and an end point, and spans data containing one or more wall span objects, each wall span object representing a structural feature along the wall segment.

[0055] In some aspects, the techniques described herein relate to a system, wherein each wall span object includes: a span type indicating a type of the structural feature, a percentage start value representing a starting position of the structural feature as a percentage of a length of the wall linestring, and a percentage end value representing an ending position of the structural feature as a percentage of the length of the wall linestring.

[0056] In some aspects, the techniques described herein relate to a property assessment apparatus for wildfire risk assessment, the apparatus including: one or more microprocessors; a display in communication with the one or more microprocessors; and non-transitory computer memory in communication with the one or more microprocessors, the non-transitory computer memory storing: a property features data structure for a target property including a target building structure, the property features data structure including: an info section storing property metadata including an address of the target property and risk data aggregated from a plurality of authoritative sources, structures data storing one or more structure objects, each structure object including a structure shape defined as a polygon in geospatial coordinates, a structure type field, walls data, and a roof object, trees data storing one or more tree objects, each tree object including a tree shape defined as a polygon, a height value, a canopy radius value, a breast height diameter value, a species designation, and trim information, an elevation section storing a grid of elevation values with geographic bounds, and other shapes data storing one or more other shape objects representing miscellaneous property features; and computer-readable instructions that, when executed by the one or more microprocessors, cause the one or more microprocessors to: A) display, onAttorney Docket No. FWIG-003W001 the display, a geospatially registered rendering of the target property based on the property features data structure; B) receive user input representing a modification to the property features data structure; C) generate a differential change record representing the modification, the differential change record including a path specifying a hierarchical location within the property features data structure; D) apply the differential change record to the property features data structure to produce an updated property features data structure; and E) provide the updated property features data structure to a property ignition model for calculation of wildfire risk.

[0057] In some aspects, the techniques described herein relate to an apparatus, wherein the computer-readable instructions, when executed by the one or more microprocessors, further cause the one or more microprocessors to: transform geometric data from WGS84 / EPSG:4326 geographic coordinates to EPSG:3857 projected coordinates for display on the display and for user interaction, and apply an inverse transformation to convert edited coordinates from EPSG:3857 projected coordinates back to WGS84 / EPSG:4326 geographic coordinates when storing changes to the property features data structure.

[0058] In some aspects, the techniques described herein relate to an apparatus, wherein the property features data structure further includes an images section storing image metadata, the image metadata including: satellite imagery metadata including geographic bounds and resolution, and a reference to a three-dimensional model of the target building structure generated from wrap-around video using structure-from-motion techniques.

[0059] In some aspects, the techniques described herein relate to an apparatus, wherein the three-dimensional model is stored in a Gaussian splatting format.

[0060] In some aspects, the techniques described herein relate to an apparatus, wherein each other shape object of the other shapes data is classified as one of: a standalone feature having a position independent of the target building structure, or an attached feature having a position defined relative to the target building structure.

[0061] In some aspects, the techniques described herein relate to a system for aligning a three-dimensional reconstruction of a building structure with geospatial reference data, the system including: a data store configured to store: image data representing a three-dimensional reconstruction of the building structure, the three-dimensional reconstruction being generated from wrap-around video of the building structure using structure-from-motion techniques, and reference data including a geospatially calibrated polygonAttorney Docket No. FWIG-003W001 representing a footprint of the building structure; an alignment engine implemented by one or more processors in communication with the data store, the alignment engine configured to: A) determine an estimated perimeter of the building structure from the three-dimensional reconstruction; B) apply an iterative closest point technique to iteratively minimize a distance between the estimated perimeter and the geospatially calibrated polygon by applying successive rigid transformations, the successive rigid transformations including rotation, translation, and uniform scaling; C) determine that alignment is successful when the estimated perimeter, after application of the successive rigid transformations, falls within a threshold distance of the geospatially calibrated polygon; and D) store alignment data representing a final transformation resulting from the successive rigid transformations; and a virtual inspection engine in communication with the data store, the virtual inspection engine configured to display the three-dimensional reconstruction using the alignment data such that structural features identified in the three-dimensional reconstruction are geospatially registered to the geospatially calibrated polygon.

[0062] In some aspects, the techniques described herein relate to a system, further including: a mobile device including an inertial measurement unit, the mobile device being configured to capture the wrap-around video of the building structure, wherein the alignment engine is further configured to receive time-synchronized inertial measurement unit data from the mobile device, the inertial measurement unit data including magnetometer readings for orientation, gyroscope data for angular velocity, and accelerometer data for linear acceleration.

[0063] In some aspects, the techniques described herein relate to a system, wherein the alignment engine is further configured to use the inertial measurement unit data to estimate a camera pose and a camera trajectory during capture of the wrap-around video without requiring manually identified ground control points.

[0064] In some aspects, the techniques described herein relate to a system, wherein the uniform scaling of the successive rigid transformations accounts for systematic scale errors in the three-dimensional reconstruction caused by focal length variations or scale drift in sequential feature tracking.

[0065] U.S. Pat. Pub. 2023 / 0023808, titled “System and Method for Wildfire Risk Assessment, Mitigation and Monitoring for Building Structures,” filed on July 12, 2022; U.S. Provisional Application No. 63 / 677,202, titled “Dimensional Geometric Modeling UsingAttorney Docket No. FWIG-003W001 Aerial and Ground Imagery,” filed on July 30, 2024; PCT / US25 / 39951, filed on July 30, 2025; U.S. Provisional Application No. 63 / 716,443, titled “System and Method for Machine-Learning Based Feature Extraction in Sparse Point Cloud Reconstructions,” filed on November 5, 2024; and PCT / US25 / 54231, filed on November 5, 2025, is each hereby incorporated by reference in its entirety for all purposes.

[0066] It should be appreciated that all combinations of the foregoing concepts and additional concepts discussed in greater detail below (provided such concepts are not mutually inconsistent) are contemplated as being part of the inventive subject matter disclosed herein. In particular, all combinations of claimed subject matter appearing at the end of this disclosure are contemplated as being part of the inventive subject matter disclosed herein. It should also be appreciated that terminology explicitly employed herein that also may appear in any disclosure incorporated by reference should be accorded a meaning most consistent with the particular concepts disclosed herein.BRIEF DESCRIPTION OF THE DRAWINGS

[0067] The skilled artisan will understand that the drawings primarily are for illustrative purposes and are not intended to limit the scope of the inventive subject matter described herein. The drawings are not necessarily to scale; in some instances, various aspects of the inventive subject matter disclosed herein may be shown exaggerated or enlarged in the drawings to facilitate an understanding of different features. In the drawings, like reference characters generally refer to like features (e.g., functionally similar and / or structurally similar elements).

[0068] For a fuller understanding of the nature and advantages of the concepts disclosed herein, reference is made to the detailed description and the accompanying drawings.

[0069] Fig. l is a flow chart of a computer-implemented method for generating a three-dimensional rendering of a building structure according to one or more embodiments.

[0070] Fig. 2 shows an example of a 3D sparse point cloud reconstruction of a building structure.

[0071] Fig. 3 is a flow chart of a method for determining an estimated perimeter of a building structure using a 3D sparse point cloud reconstruction.Attorney Docket No. FWIG-003W001

[0072] Fig. 4 shows an example of an estimated perimeter determined from projecting high-density vertical columns of points onto a ground plane according to one or more embodiments.

[0073] Fig. 5 shows a simplified example of a volume-rendered scaled 3D sparse point cloud reconstruction of a building structure according to one or more embodiments.

[0074] Fig. 6 is a flow chart of a computer-implemented method for performing a virtual wildfire inspection of a building structure according to one or more embodiments.

[0075] Fig. 7A is an overhead view of the building structure and an observation / inspection point (e.g., of a virtual wildfire inspector) according to one or more embodiments.

[0076] Fig. 7B is a simplified view of a display screen that displays a viewable portion of the building structure corresponding to or representing the position, field-of-view (FOV), and orientation of the observation / inspection point shown in Fig. 7A.

[0077] Fig. 8A is an overhead view of the building structure showing an example updated position and updated FOV of the observation / inspection point compared to the position and FOV of the observation / inspection point shown in Fig. 7A.

[0078] Fig. 8B is a simplified view of an updated viewable portion of the building structure on the display screen corresponding to or representing the position, FOV, and orientation of the observation / inspection point shown in Fig. 8A.

[0079] Fig. 9A is an overhead view of the building structure showing an example updated orientation of the observation / inspection point compared to the orientation of the observation / inspection point shown in Figs. 8A and 9A.

[0080] Fig. 9B is a simplified view of an updated viewable portion of the building structure on the display screen corresponding to or representing the position, FOV, and orientation of the observation / inspection point shown in Fig. 9A.

[0081] Fig. 10 is a flow chart of a computer-implemented method for performing a virtual wildfire inspection of a building structure according to one or more embodiments.

[0082] Fig. 11 shows the updated viewable portion of the building structure in Fig. 9B where the wood pile has been tagged as a tagged object.

[0083] Fig. 12 is a block diagram representing the data for a tagged object that can be stored.Attorney Docket No. FWIG-003W001

[0084] Fig. 13 is a flow chart of a computer-implemented method for performing a virtual wildfire inspection of a building structure according to one or more embodiments.

[0085] Fig. 14A shows an example of a virtual inspection report that includes a table.

[0086] Fig. 14B shows an example of a virtual inspection report that includes one or more sections.

[0087] Fig. 15 is a flow chart of a computer-implemented method for detecting and / or segmenting structural features in digital images according to one or more embodiments.

[0088] Fig. 16 is an example 3D sparse point cloud rendering of a building structure that includes masked colors for the structural features according to one or more embodiments

[0089] Fig. 17 is an example 3D sparse point cloud rendering of a building structure that includes bounding boxes corresponding to masked structural features according to one or more embodiments.

[0090] Fig. 18 is a simplified view of a display screen that displays a viewable portion of a building structure where each type of structural feature is masked in a different color according to one or more embodiments.

[0091] Fig. 19 is a flow chart of a method for remotely monitoring a building structure for compliance of wildfire mitigation actions.

[0092] Fig. 20 is a flow chart of a method for performing a virtual fire inspection according to one or more embodiments.

[0093] Fig. 21 shows an example overhead image of a property including a building structure according to one or more embodiments.

[0094] Fig. 22 is a flow chart of a computer-implemented method for guiding a person such as a service provider, via a mobile computing device carried by or associated with the person, to perform wildfire mitigation actions for a building structure according to one or more embodiments.

[0095] Fig. 23 shows an example viewable portion of a building structure, as shown on a display, that includes several tagged objects.

[0096] Fig. 24 shows an example viewable portion of a three-dimensional rendering of a building structure, as shown on a display, according to one or more embodiments.Attorney Docket No. FWIG-003W001

[0097] Fig. 25 shows an example viewable portion of a three-dimensional rendering of a building structure, as shown on a display, according to one or more embodiments.

[0098] Fig. 26 is a block diagram of a system for performing a virtual inspection of a building structure according to one or more embodiments.

[0099] Fig. 27 shows an example viewable portion of a three-dimensional rendering of a building structure, as shown on a display, according to one or more embodiments.

[0100] Fig. 28 shows an example viewable portion of a three-dimensional rendering of a building structure, as shown on a display, according to one or more embodiments.

[0101] Fig. 29 is a block diagram of the image processing engine of Fig. 26 according to one or more embodiments.

[0102] Fig. 30 is a block diagram of the annotation engine of Fig. 26 according to one or more embodiments.

[0103] Fig. 31 is a block diagram of the wildfire ignition risk modelling engine of Fig. 26 according to one or more embodiments.

[0104] Fig. 32 is a block diagram of a system and method for performing virtual inspections of a target building structure to determine a risk of loss to the building structure due to a wildfire according to one or more embodiments.

[0105] Fig. 33 is a block diagram of the image processor of Fig. 32 according to one or more embodiments.

[0106] Fig. 34 is a block diagram of the virtual inspection engine of Fig. 32 according to one or more embodiments.

[0107] Fig. 35 A shows a geospatially calibrated aerial image of a target property including a target building structure.

[0108] Fig. 35B shows the geospatially calibrated aerial image of Fig. 35 A overlaid with a portion of a volume-rendered reconstruction of the target building structure, according to one or more embodiments.

[0109] Figs. 36A shows a simultaneous display of an outlined two-dimensional projection of a volumetrically rendered 3D reconstruction of a target building structure and an outlined two-dimensional projection of the geospatial footprint or ground plane in a misaligned state.Attorney Docket No. FWIG-003W001

[0110] Figs. 36B shows a simultaneous display of an outlined two-dimensional projection of a volumetrically rendered 3D reconstruction of a target building structure and an outlined two-dimensional projection of the geospatial footprint or ground plane in an aligned state.

[0111] Fig. 37 is a block diagram of the post processor of Fig. 32 according to one or more embodiments.

[0112] Fig. 38 is a flow chart of a method for generating a geospatially calibrated overhead rendering of a target property according to one or more embodiments.

[0113] Fig. 39 is a block diagram of a virtual inspection apparatus according to one or more embodiments.

[0114] Fig. 40 is a block diagram of a property features data structure according to some embodiments.

[0115] Fig. 41 is a block diagram of a structure data model including wall span representation according to some embodiments.

[0116] Fig. 42 is a block diagram of a tree data model including trim information according to some embodiments.

[0117] Fig. 43 is a flow diagram of a property features data flow from initial generation through correction and processing according to some embodiments.

[0118] Fig. 44 is a diagram illustrating geospatially located wall spans using percentagebased representation according to some embodiments.

[0119] Fig. 45 is a block diagram of an other shapes data model for supplementary property features according to some embodiments.

[0120] Fig. 46 illustrates a networked computing system according to one or more embodiments.

[0121] Fig. 47 is an example block diagram of a computing device that may incorporate one or more embodiments of the present disclosure.DETAILED DESCRIPTION

[0122] Images representing all outside walls of a building structure, including any structural features on / in the outside walls, are used to generate a three-dimensional (3D) computer rendering of the building structure. In one example implementation, respective images ofAttorney Docket No. FWIG-003W001 outside walls of a building structure may be acquired as a video of the exterior of the building structure (e.g., taken with one or more cameras of a smartphone or other device held by a person and directed at the exterior of the building structure as the person walks around the exterior of the building structure). The images are first transformed into a 3D sparse point cloud to form a 3D sparse point cloud reconstruction of the building structure. Such a 3D sparse point cloud reconstruction of the building structure is then volumetrically-rendered to form the 3D rendering. In one aspect, the 3D sparse point cloud is geospatially registered and geospatially calibrated. In other aspects, a viewable portion of the 3D rendering can be displayed according to a position, a field-of-view, and / or an orientation of an observation / inspection point (e.g., corresponding to a location of a virtual wildfire inspector or a virtual camera) relative to the building structure. The viewable portion of the 3D rendering can be iteratively updated as the position, the field-of-view, and / or the orientation of the observation / inspection point changes.

[0123] A geospatially calibrated two-dimensional (2D) rendering of an overhead view of a property that includes the building structure can be generated and displayed. The geospatially calibrated 2D rendering can be displayed separately from the 3D rendering or overlaid on the 3D rendering. The geospatially calibrated 2D rendering can be generated from a geospatially calibrated aerial image that includes the property. The perimeter of the building structure can be detected in the geospatially calibrated aerial image to produce a geospatially calibrated polygon that defines the footprint of the building structure. Large vegetation such as trees can be detected in the geospatially calibrated aerial image and geospatially calibrated bounding regions can define the detected large vegetation.

[0124] Objects such as structural features and / or fuel sources represented in the viewable portion of the 2D and / or the 3D rendering can be tagged. A geospatial position and a geospatial size and a respective image position and a respective image size of a tagged object is automatically stored in a property features data store with each tag. In one or more embodiments, an image of the tagged object can be automatically generated from the 3D rendering and stored in the property features data store. In one or more embodiments, an image representing a position of the tagged object relative to the building structure can be automatically generated and stored, in a property features data stored.

[0125] The tagged objects can represent planned wildfire mitigation actions and / or actions to address insurance policy violations, for example to be used by a service provider who may not be familiar with the property and building structure. The wildfire mitigation actionsAttorney Docket No. FWIG-003W001 and / or actions to address insurance policy violations, performed by the service provider, can be tracked in a mobile application using images acquired by the service provider and / or geospatial position data of a mobile device held by the service provider. After service is complete, the service provider can capture a new set of images representing all outside walls of the building structure that can be used to generate a new 3D computer rendering of the building structure that can be used for virtual inspection to confirm that the planned wildfire mitigation actions and / or actions to address insurance policy violations were completed.

[0126] Fig. 1 is a flow chart of a computer-implemented method 10 for generating a three-dimensional rendering of a building structure according to one or more embodiments.Method 10 can be performed by one or more hardware-based processors in a computer for example by executing computer-readable instructions stored in the non-volatile memory of the computer.

[0127] In optional step 100, image data representing digital images that collectively represent or show all outside walls of a building structure are received. Though the digital images collectively represent all outside walls of a building structure, any one digital image may only represent one outside wall on the building structure or even a portion (e.g., a segment) of one outside wall on the building structure. In one or more embodiments, some or all of the digital images can have a field-of-view sufficient to show the entire height (vertical span) of the structure.

[0128] The digital images can comprise or represent digital photographs (e.g., digital photo data) captured with a portable (e.g., handheld) digital camera, such as a digital camera in a portable computer such as smartphone, a tablet, or smart glasses (e.g., Meta glasses available from Meta Platforms Inc.) or a dedicated digital camera. Additionally or alternatively, the digital images can comprise or represent digital video(s) (e.g., digital video data) and / or sampled digital video images (e.g., sampled digital video image data) from one or more digital videos. The digital video(s) can be captured with a portable (e.g., handheld) digital video camera such as a digital video camera in a portable computer such as smartphone, a tablet, or smart glasses (e.g., Meta glasses available from Meta Platforms Inc.) or a dedicated digital video camera. Additionally or alternatively, some or all of the digital photographs and / or some or all of the digital videos can be captured with a digital camera and / or a digital video camera, respectively, that can be included in or mounted on a drone or robot. In one or more embodiments, the digital video(s) can be created while panning and / or physically moving the digital video camera up and / or along one or more sides and / or outside walls ofAttorney Docket No. FWIG-003W001 the building structure, such as in a multi-story structure, to capture all or substantially all of the structural features on the outside walls. In one or more embodiments, a digital video can represent all the outside walls of a building structure while the digital video camera is moved in at least onecomplete loop (e.g., only one loop, one loop and at least a portion of a second loop, or multiple loops) about the perimeter of the building structure. The digital video camera can be held by a person who walks in at least one loop about the perimeter of the building structure to capture the digital video. Alternatively, the digital video camera can placed or mounted on / in a drone or robot that moves in at least one loop about the perimeter of the building structure to capture the digital video.

[0129] The building structure can be or comprise a residential structure such as a house, a condominium, a townhouse, or a residential building, or can be or comprise a commercial structure such as an office building, a retail store, or a factory. The building structure is located on property that can include objects such as natural objects (e.g., trees, shrubs, grass, rocks, and / or other natural objects) and / or artificial (man-made) objects (e.g., vehicles, secondary structures such as a shed and / or a pool house, furniture, fences, signs, industrial storage containers, pavement, and / or other artificial objects).

[0130] The digital images can be received or provided from computer memory (e.g., nonvolatile memory) in the computer or in another computer in communication the computer. In one or more embodiments, the digital images are received directly or indirectly from a portable computer, such as a smartphone, that includes a digital camera to capture the digital images.

[0131] In step 101, a plurality of digital images that collectively represent all outside walls of a building structure are transformed into a three-dimensional (3D) sparse point cloud representation (“reconstruction”) of the building structure.

[0132] By way of example, a sparse point cloud of the digital images can be created using sparse point cloud reconstruction and / or Structure-from-Motion methods such as COLMAP or Hierarchical Localization. COLMAP is a general-purpose Structure-from-Motion and Multi-View Stereo pipeline, created by Johannes L. Schoenberger, available at https: / / colmap.github.io / , which is hereby incorporated by reference. Hierarchical Localization can be performed using a hierarchical localization toolbox, a modular toolbox that leverages image retrieval and feature matching, created by Paul-Edouard Sarlin,Attorney Docket No. FWIG-003W001 available at https: / / github.com / cvg / Hierarchical-Localization, which is hereby incorporated by reference.

[0133] The inventors have recognized and appreciated, however, that various aspects of the digital images used to create a sparse point cloud can significantly improve the integrity and utility of the sparse point cloud for ultimately generating a 3D rendering of the building structure to facilitate a fire inspection, as set forth in further detail below.

[0134] For example, the respective digital images (e.g., or video) of the outside walls of the building structure should contain as much of the vertical span of a given outside wall of the building structure as possible (and preferably the entire vertical span of the outside wall). Accordingly, when the digital images are acquired (e.g., via a camera, smartphone or other image acquisition device), as much of the vertical span of the outside wall as possible should be kept within the viewing frame of the device (and optionally some additional space above and below the vertical span of the outside wall may be kept in the viewing frame).

[0135] Additionally, during image acquisition, quick movements or panning of the camera, smartphone or other image acquisition device should be avoided, so as to mitigate motion blur across multiple digital images. Furthermore, the digital images should be acquired by making at least one entire loop around the perimeter of the building structure so as to acquire digital images of all portions of all outside walls. In some implementations, at least one entire loop must be made around the perimeter of the building structure to ensure sufficient coverage of the digital images to effectively generate a sparse point cloud. In one aspect, the end point of a loop around the building may continue past the start point of the loop such that an ending portion of the loop overlaps with a starting portion of the loop; this ensures that all of the outside walls of the building structure are represented in the digital images (even if some portions of a given outside wall are repeated in multiple images). In another aspect, if image processing power permits, multiple loops can be made around the perimeter of the building structure to acquire the digital images used to generate the sparse point cloud. In yet another aspect, if the camera, smartphone or other image acquisition device is carried / operated by a person, the systems and methods disclosed herein may provide to the person, via the device, one or more prompts instructing the user to: 1) capture the entire vertical span of the building structure during image acquisition; 2) avoid quick movements or panning with the device during image acquisition; and 3) make at least one full loop around the perimeter of the building structure while acquiring the digital images.Attorney Docket No. FWIG-003W001

[0136] In one or more embodiments, the digital images can and / or should include at least some of the property adjacent to the building structure. The property adjacent to the building structure (e.g., within a range of about 30 feet from the outside walls) can include fuel-source objects that can contribute to the wildfire ignition risk of the building structure. Examples of such fuel-source objects can include vegetation, such as trees, shrubs, foundation plantings, and / or other plants, and / or other objects such as outdoor furniture, wood piles, leaf piles, raised beds, fuel tanks, fences, secondary structures, and / or other fuel sources. In an aspect, the systems and methods disclosed herein may provide to the person, via the device, one or more additional prompts instructing the user to: 4) capture the property surrounding the building structure during image acquisition; 5) stand a predetermined distance (e.g., 30-50 feet) from the building structure to capture the nearfield property during image acquisition; and / or 6) perform a 360-degree pan of the digital camera (or digital video camera) at least once on each side of the building structure.

[0137] Fig. 2 shows an example of a 3D sparse point cloud reconstruction 20 of a building structure. The 3D sparse point cloud reconstruction 20 includes a sparse point cloud 200. Each point 210 in the sparse point cloud 200 has a respective color 215 and voxel location according to the digital images. In one aspect, the sparse point cloud 200 is “non-dimensional” in that a distance between any two points in the cloud does not represent any physical value (such as a distance between physical elements of the building structure or its environs). Fig. 2 also illustrates an estimated ground plane 410 shown with respect to the sparse point cloud 200 of the spare point cloud reconstruction 20. Generation of the estimated ground plane 410 is discussed in detail below.

[0138] In acquiring the digital images used to generate the sparse point cloud 200, it should be appreciated that, in one aspect, the orientation of the image acquisition device with respect to the building structure during at least the first digital image (or first frame of a video) acquired of an outside wall of the building structure determines a three-dimensional (3D) coordinate system or “grid” of voxels on which the sparse point cloud is generated. The respective x, y and z dimensions of each voxel do not necessarily correspond to any actual physical dimension, but instead represent some unit of resolution between minimum and maximum values for the respective dimensions in the grid space. The sparse point cloud that is generated based on processing of the subsequent digital images is going to be constructed with reference to the 3D coordinate system established by at least the first digital image. Accordingly, if the image acquisition device is tilted with respect to the outside wall of theAttorney Docket No. FWIG-003W001 building structure during acquisition of at least the first digital image (e.g., an optical axis of the image acquisition device is not essentially normal to the building structure, or not essentially parallel to the ground surface below the building structure), the sparse point cloud that is generated based on processing of the subsequent digital images is also going to be similarly tilted with respect to the perspective of the actual building structure and the ground surface below the building structure. In view of the foregoing, respective additional steps of the method 10 shown in Fig. 1 and discussed further below apply inventive processing concepts to account for an arbitrary tilting of the 3D coordinate system on which the sparse point cloud initially is generated, to in turn generate a geospatially-calibrated spare point cloud that accurately represents the building structure (and in some instances the immediate surroundings of the building structure).

[0139] To this end, and with reference again to Fig. 1, in step 102, the 3D sparse point cloud 200 generated in step 101 and shown in Fig. 2 is further processed to determine an estimated ground plane, and in step 103 an estimated perimeter of the building structure is determined based on the estimated ground plane. In one or more embodiments, steps 102 and 103 of Fig.1 can be performed according to steps 301-303 in Fig. 3. Turning now to Fig. 3, in step 301 the 3D grid on which the sparse point cloud 200 is generated is divided into vertical columns, in which the z dimension (or height dimension) of each vertical column is taken as infinite (or the maximum available value in the z dimension), and in which the x and y dimensions of each vertical column are the same (e.g., to form a square, using an equal number of voxels; in one example implementation, the number of voxels corresponds to approximately one foot by one foot in physical distance). Density-based scanning is then performed on each vertical column to determine the lowest point of the sparse point cloud 200 in each vertical column.

[0140] In step 302, a plane is then fit across the lowest points of the respective vertical columns to define the estimated ground plane 410 shown in Fig. 2. Once the ground plane is defined, in step 303 all the points in the 3D sparse point cloud are projected onto that plane. In step 303, the projected points correspond to the estimated perimeter of the building structure. Fig. 4 shows an example of the projected 3D sparse points resulting in the estimated perimeter 400 (i.e., all the points in the 3D sparse point cloud projected onto the ground plane 410 shown in Fig. 2.

[0141] Fig. 4 also shows a geospatially calibrated polygon 420 that represents the geospatial perimeter or geospatial footprint of the building structure. In one or more embodiments, the geospatially calibrated polygon 420 can be provided by and / or received from a third party.Attorney Docket No. FWIG-003W001 Additionally or alternatively, a trained machine-learning (ML) structure footprint detection model can be used to detect the structure footprint in one or more overhead, aerial, and / or satellite images of the building structure. The trained ML structure footprint detection model can be trained with sample overhead, aerial, and / or satellite images and known footprints of building structures.

[0142] Returning for the moment to the method 10 shown in Fig. 1, in step 104 the estimated perimeter 400 is registered and / or aligned with a geospatially calibrated polygon 420 that represents the geospatial perimeter or geospatial footprint of the building structure. In one example implementation, the estimated perimeter 400 and the geospatially calibrated polygon 420 can be spatially aligned by rotating, translating, and scaling the estimated ground plane 410 to create a best overlay fit of the estimated perimeter 400 and the geospatially calibrated structure perimeter 420 as determined by iterative closest point (ICP) alignment or another method. The axes and values shown in Fig. 4 are provided as examples and are not intended to be limiting.

[0143] As a result of alignment and / or registration, the estimated perimeter 400 is placed in a known or common coordinate space of the geospatially calibrated polygon 420.

[0144] Returning to Fig. 1, in step 105 the 3D sparse point cloud 200 can be scaled according to the alignment / registration performed in step 104 so as to generate a scaled sparse point cloud reconstruction. When the estimated polygon 420 is rotated and / or translated during alignment / registration, the 3D locations (e.g., voxel locations) of the points in the 3D sparse point cloud 200 are rotated and / or translated, respectively, relative to their original locations. For example, if the rotational orientation of the estimated polygon 420 is changed as a result of alignment / registration, the corresponding location (e.g., voxel location) of each point in the 3D sparse point cloud 200 is rotationally varied accordingly. Likewise, if the estimated polygon 420 is translated parallel to the ground plane 410 as a result of alignment / registration, the corresponding location (e.g., voxel location) of each point in the 3D sparse point cloud 200 is translated laterally accordingly.

[0145] If the size of the estimated polygon 420 is changed as a result of alignment / registration, the 3D locations (e.g., voxel locations) of the points in the 3D sparse point cloud 200 are varied accordingly, relative to their original locations. For example, if the size of the estimated polygon 420 is increased as a result of alignment / registration, then a given point in the 3D sparse point cloud 200 (e.g., representing a portion of a window) mayAttorney Docket No. FWIG-003W001 be located further away from the ground plane and / or laterally further away from a centroid of the estimated polygon than that point was originally (prior to the increase in size of the estimated polygon). Likewise, if the size of the estimated polygon 420 is decreased as a result of alignment / registration, then a given point in the 3D sparse point cloud 200 (e.g., representing a portion of a window) may be located closer to the ground plane 410 and / or laterally closer to a centroid of the estimated polygon than that point was originally (prior to the decrease in size of the estimated polygon).

[0146] As a result of scaling in step 105 of Fig. 1, the 3D locations (e.g., voxel locations) of the points in the 3D sparse point cloud 200 are in the known or common coordinate space of the geospatially calibrated polygon 500. The voxel location of each point in the 3D sparse point cloud 200 has a corresponding geospatial position that can be stored in non-volatile memory.

[0147] Returning to Fig. 1, in step 106 a 3D computer rendering of the building structure is generated. The 3D rendering can comprise or can be the scaled 3D sparse point cloud reconstruction discussed above in connection with step 105. Alternatively, the scaled 3D sparse point cloud reconstruction can be a “volumetrically-rendered reconstruction” to generate a more realistic 3D rendering which can be or can be close to photorealistic in one or more embodiments. In one aspect, volumetric rendering can be used to “paint” the colors represented in the respective points of the scaled 3D sparse point cloud so as to generate a volumetrically-rendered reconstruction.

[0148] In one or more embodiments, the scaled 3D sparse point cloud reconstruction can be volumetrically-rendered using Gaussian Splatting. Gaussian Splatting can be performed as described in “3D Gaussian Splatting for Real-Time Radiance Field Rendering” by Bernhard Kerbl et al, ACM Trans. Graph., Vol. 42, No. 4, published August 2023 and / or at https: / / repo-sam.inria.fr / fungraph / 3d-gaussian-splatting / , which are hereby incorporated by reference.

[0149] In one or more other embodiments, the scaled 3D sparse point cloud reconstruction can be volumetrically-rendered using Neural Radiance Field (NeRF). NeRF can be performed as described in “NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis” by Ben Mildenhall et al., European Conference on Computer Vision (ECCV) Oral Presentation, arXiv:2003.08934v2, August 3, 2020 and / or at https: / / www.matthewtancik.com / nerf, which are hereby incorporated by reference.Attorney Docket No. FWIG-003W001

[0150] Fig. 5 shows a simplified example of a volume-rendered scaled 3D sparse point cloud reconstruction (volume-rendered reconstruction 60) of a building structure 600 according to one or more embodiments. The volume-rendered reconstruction 60 includes the outside walls 610 of the building structure 600 and structural features 620 defined on, in, and / or connected to the outside walls 610. Examples of structural features 620 include windows 622, doors 624, decks 626, balconies, vents 628 (e.g., air vents, laundry vents, and / or other vents), soffits, and / or garage doors 630. Though the roof 640 may not be viewable in the digital images representing the outside walls of the building structure (e.g., that are transformed in step 101 to a sparse point cloud representation), the volume-rendered reconstruction 60 can include a roof 640 (though the rendering of the roof 640 may differ in appearance from the roof of the actual building structure).

[0151] The volume-rendered reconstruction 60 can include additional objects 650 that are viewable in the digital images representing the outside walls of the building structure (e.g., that are transformed in step 101 to a sparse point cloud representation). Examples of these objects 650 include patio furniture 652, foundation plantings 654, fences 656, raised beds 658, and / or wood piles 660. One, some, or all of the additional objects 650 may represent fuel sources that can increase the wildfire ignition risk of the building structure 600.

[0152] Fig. 6 is a flow chart of a computer-implemented method 70 for performing a virtual wildfire inspection of a building structure according to one or more embodiments. Method 70 can be performed by one or more hardware-based processors in a computer for example by executing computer-readable instructions stored in the non-volatile memory of the computer.

[0153] In step 701, only a portion of the 3D rendering (e.g., the volume-rendered reconstruction 60) is displayed on a display screen. The display screen can be coupled to the same computer that generated the 3D rendering (e.g., in step 106 of Fig. 1) such that method 10 and method 70 are performed on the same computer. Alternatively, the display screen can be coupled to a different computer that generated the 3D rendering, such as in a client-server relationship. The portion of the 3D rendering that is displayed can correspond to or represent a field of view (FOV), orientation, and position of an observation / inspection point (e.g., a location of a virtual wildfire inspector) relative to the building structure. Image data for generating the 3D rendering can be included in property features data stored locally or in a property features data store, or in a unique link provided in the property features data.Attorney Docket No. FWIG-003W001

[0154] Fig. 7A is an overhead view of the building structure 600 showing an example observation / inspection point 800 (e.g., of a virtual wildfire inspector) having a position 810, a FOV 820, and an orientation 830 relative to the building structure 600. In Fig. 7A, the position 810 of the observation / inspection point 800 is in the middle of outside wall 610A. The orientation 830 of the observation / inspection point 800 is directed towards the middle of the outside wall 610A. The FOV 820 can be defined by a FOV angle 822.

[0155] Fig. 7B is a simplified view of a display screen 80 that displays a viewable portion 840 of the building structure 600 corresponding to or representing the position 810, FOV 820, and orientation 830 of the observation / inspection point 800 in Fig. 7A. As shown in Fig. 7B, the viewable portion 840 includes outside wall 610A but does not include outside walls 610B-D. The structural features 620 (e.g., windows 622, door 624, and garage door 630) defined on, in, and / or connected to the outside wall 610A are displayed in the viewable portion 840. To the extent any additional objects 650, that were captured in the digital images transformed in step 101, are located in front of the outside wall 610A and within the FOV 820, those objects can be shown as part of the viewable portion 840.

[0156] Returning to Fig. 6, in step 702 one or more user inputs is / are received to change the position, FOV, and / or orientation of the observation / inspection point 800. The user inputs can be provided by gesture (e.g., on a touch screen), a mouse, a keyboard, voice command, and / or other user input. In one or more embodiments, the user input can be provided through a graphical user interface that can be configured to represent movement of the observation / inspection point (e.g., the location of a virtual wildfire inspector) relative to the building structure.

[0157] In one or more embodiments, the virtual wildfire inspector and / or the position of the virtual wildfire inspector relative to the building structure can be displayed on the display screen. In one or more other embodiments, the virtual wildfire inspector and / or the position of the virtual wildfire inspector relative to the building structure is / are not shown on the display screen.

[0158] In step 703, an updated portion of the 3D rendering is displayed on the display screen. The updated portion corresponds to or represents the new / updated FOV, orientation, and position of the observation / inspection point relative to the building structure (e.g., according to the user input(s) received in step 702).Attorney Docket No. FWIG-003W001

[0159] Fig. 8A is an overhead view of the building structure 600 showing an example updated position 910 and updated FOV 920 of the observation / inspection point 800 compared to the position 810 and FOV 820 of the observation / inspection point 800 shown in Fig. 7A. The observation / inspection point 800 has moved closer to the outside wall 610A in front of the door 624. The updated position 910 and the updated FOV 920 can be updated, relative to the initial position 810 and initial FOV 820 of the observation / inspection point 800 shown in Fig. 7A, in response to one or more user inputs received in step 702. The updated FOV 920 can be defined by an updated FOV angle 922. The size or scope of the updated FOV 920 (e.g., the updated FOV angle 922) is smaller than the size or scope of the FOV 820 (e.g., than the FOV angle 822). In one or more other embodiments, the size or scope of the updated FOV 920 (e.g., the updated FOV angle 922) can be larger than the size or scope of the FOV 820 (e.g., than the FOV angle 822).

[0160] Fig. 8B is a simplified view of an updated viewable portion 940 of the building structure 600 on the display screen 80. The updated viewable portion 940 corresponds or represents the updated position 910 and the updated FOV 920.

[0161] In one or more embodiments, steps 702 and 703 can be repeated in a loop 704 to iteratively update the position, FOV, and / or orientation of the observation / inspection point and to iteratively update the displayed portion of the 3D rendering. Each update to the displayed portion of the 3D rendering (step 703) corresponds to or represents a respective update to the position, FOV, and / or orientation of the observation / inspection point (in response to the user input(s) received in step 702).

[0162] Fig. 9A is an overhead view of the building structure 600 showing an example updated orientation 1030 of the observation / inspection point 800 compared to orientation 830 of the observation / inspection point 800 shown in Figs. 8A and 9A. The observation / inspection point 800 has turned to the right in Fig. 9A, compared to Figs. 8A and 9A, such that a portion of the door 624, one of the windows 620, and the wood pile 660 are within an updated FOV 1020 (having an updated FOV angle 1022) of the observation / inspection point 800. The observation / inspection point 800 is in the same position 910 in Fig. 9A and Fig. 10. The updated orientation 1030 is determined according to the user input(s) received in step 702. The updated FOV angle 1022 can be the same as the updated FOV angle 922.Attorney Docket No. FWIG-003W001

[0163] Fig. 9B is a simplified view of an updated viewable portion 1040 of the building structure 600 on the display screen 80. The updated viewable portion 1040 corresponds or represents the updated orientation 1030 and the updated FOV 1020.

[0164] Fig. 10 is a flow chart of a computer-implemented method 1100 for performing a virtual wildfire inspection of a building structure according to one or more embodiments. Method 1100 can be performed by one or more hardware-based processors in a computer for example by executing computer-readable instructions stored in the non-volatile memory of the computer. In one or more embodiments, method 1100 can be performed to tag objects in a 3D rendering (e.g., a volume-rendered reconstruction) and associated mitigation actions to be performed.

[0165] Method 1100 adopts some of the same preliminary steps as shown in method 70 of Fig. 6, and includes additional steps beginning with step 1104 of receiving a user input to tag an object. The tagged object can represent a policy violation (e.g., as specified by one or more terms of a wildfire insurance policy). Additionally or alternatively, the tagged object can represent a fuel source that may be located too close to (e.g., within a predetermined distance of) the building structure, which may increase the wildfire risk to the building structure. The wildfire risk to the building structure can be determined using a property ignition model (PIM), for example as described in U.S. Application Publication No.2023 / 0023808, titled “System And Method For Wildfire Risk Assessment, Mitigation And Monitoring For Building Structures,” which is hereby incorporated by reference.

[0166] In optional step 1105, the state or appearance of an object (e.g., as it appears on a display of a user device) can change when the object is tagged (e.g., from an untagged state to a tagged state). With reference for the moment to Fig. 11, a symbol such as check mark can be overlaid on the tagged object 1200 to represent that the object has been tagged. In another example, the visual appearance such as the color, shading, and / or fill pattern (or other displayed visible attribute) of the tagged object 1200 can be different than that of an untagged object.

[0167] Returning to Fig. 10, in optional step 1106, one or more GUI elements for the tagged object can be displayed. The GUI element(s) can include one or more text fields, one or more drop-down lists, one or more radio buttons, and / or one or more other GUI elements. The GUI element(s) can be configured to allow a user to input information relating to the tagged object (e.g., so as to annotate the tagged object in some manner). For example, the GUI element(s)Attorney Docket No. FWIG-003W001 can include one or more object description GUI elements, one or more object category GUI elements, one or more remedial instructions GUI elements, one or more policy violation category GUI elements, one or more notes GUI elements, one or more object count GUI elements, and / or one or more object priority GUI elements.

[0168] The object description GUI element(s) can allow a user to enter or provide a description of the tagged object. The object category GUI element(s) can allow a user to enter or provide a category for the tagged object. The remedial instructions GUI element(s) can allow a user to enter or provide remedial instructions for reducing, eliminating, or mitigating the wildfire risk to the building structure caused by and / or related to the tagged object and / or for eliminating the policy violation associated with the tagged object. In one or more embodiments, the remedial instructions can indicate one or more short-term mitigation actions to protect the building structure in response to wildfire monitoring that indicates an increased wildfire risk to the region including the property on which the building structure is located. Examples of such short-term mitigation actions can include removing organic debris (e.g., leaves, sticks, etc.), applying a fire retardant (e.g., to a mulch bed and / or to vegetation such as a bush), sealing the garage doors, cleaning the gutters, and / or moving outdoor furniture and / or grills. Short-term mitigation actions can also be referred to as emergency mitigation actions.

[0169] The policy violation category GUI element(s) can allow a user to enter or provide a category for the policy violation associated with the tagged object. The notes GUI element(s) can allow a user to enter or provide notes relating to the tagged object, the remedial instructions, and / or the policy violation, and / or other notes. The object count GUI element(s) can allow a user to enter or provide the number of items associated with the tagged object. The object priority GUI element(s) can allow a user to enter or provide a priority for the tagged object and the associated mitigation task(s). The priority can be quantitative (e.g., from 1 to 10 with 10 being the highest priority and 1 being the lowest priority (or vice versa)), qualitative (e.g., high, medium, or low), or relative to the priorities of the other tagged objects. Generally speaking, any of the foregoing GUI elements relating to the tagged object may be referred to and considered as an “annotation” associated with the tagged object. The indexed location and size, in image space and geospatial space, and the annotation(s) can be stored as property features data in a property features data store, as discussed below.Attorney Docket No. FWIG-003W001

[0170] Fig. 11 shows the updated viewable portion 1040 of the building structure 600 where the wood pile 660 has been tagged (in response to user input in step 1104 of Fig. 10) as a tagged object 1200. GUI elements 1210 for the tagged object 1200 are displayed on the side of the display 80. There can be additional or fewer GUI elements 1210 (e.g., at least one GUI element 1210) in one or more embodiments.

[0171] Using the wood pile 660 as an example of a tagged object 1200, the GUI elements 1210 can include object description GUI element(s) that allow a user to enter or provide a description (e.g., wood pile), an object category (e.g., combustible material), remedial instructions (e.g., move more than 30 feet from building structure), the policy violation category (e.g., combustible material within 30 feet of building structure), notes, and / or the object count (e.g., 6 logs in wood pile).

[0172] Returning to Fig. 10, in optional step 1107 input data for the GUI element(s) displayed in step 1106 is received.

[0173] In step 1108, data for the tagged object is stored in non-volatile memory. The stored data includes a unique identifier for the tag, the data input (if any) by user for the GUI element(s) received in optional step 1107, and the geospatial location of the tagged object. The geospatial location of the tagged items is determined by the user within the geospatially aligned point cloud space. Any point specified in the aligned point cloud represents an actual geospatial location (i.e., the point cloud can include things that are not just the structure which can be tagged). The point cloud includes any object in the video frame. In one or more embodiments, the stored data for the tagged object can include an image of the tagged object and / or an image representing the location of the tagged object relative to the building structure. The data for the tagged object can be stored locally and can be synchronized with property features data in a property features data store, as discussed below.

[0174] Fig. 12 is a block diagram representing data 1300 for a tagged object 1200 that can be stored in step 1108 of Fig. 10. The data 1300 can include unique identifier data 1301, input data 1302 describing the tagged object 1200 (e.g., received in optional step 1107), geospatial location data 1303 representing the geospatial location of the tagged object 1200, tagged object image data 1304 that represents an image 1310 (e.g., a zoomed-in image) of the tagged object 1200, and / or position image data 1305 that represents an image 1320 (e.g., an overhead image or an overhead rendering) that shows the location of the tagged object 1200 relative to the building structure 600.Attorney Docket No. FWIG-003W001

[0175] The image 1310 can comprise some of or all of a viewable portion (e.g., updated viewable portion 1040) of the building structure 600 (e.g., as shown in Fig. 11) in which the tagged object 1200 is viewable. Alternatively, the image 1310 can comprise another image of the tagged object 1200 that can be generated or created from the volume-rendered reconstruction 60.

[0176] The image 1320 can comprise an annotation of the estimated polygon 420 or an annotation of the geospatially calibrated polygon 500 where the annotation indicates the position or location of the tagged object 1200 relative to the estimated polygon 420 or relative to the geospatially calibrated polygon 500, respectively.

[0177] Returning to Fig. 10, steps 702, 703, 1104, and 1108 and optional step(s) 1105, 1006, and / or 1107 can be repeated in a loop 1109 such that multiple objects can be tagged as the virtual inspector is iteratively moved with respect to the building structure (e.g., with respect to the volume-rendered reconstruction 60 of the building structure). An object may not be tagged in a given iteration through the loop 1109 (e.g., in a given iteration of the loop, the steps 1104 and 1008 can be optional, such as when no policy violations are detected).

[0178] Fig. 13 is a flow chart of a computer-implemented method 1400 for performing a virtual wildfire inspection of a building structure according to one or more embodiments. Method 1400 can be performed by one or more hardware-based processors in a computer for example by executing computer-readable instructions stored in the non-volatile memory of the computer.

[0179] In step 1401, multiple (e.g., a plurality of) portions of the 3D rendering are iteratively displayed on a display. Each portion of the 3D rendering can be displayed according to a corresponding FOV, a corresponding orientation, and a corresponding position of a virtual inspector relative to the building structure, for example according to steps 701-703 (see Fig(s). 7 and / or 11). Image data for the 3D rendering of the outside walls of the building structure can be included in property features data stored locally or in a property features data store, or in a unique link provided in the property features data.

[0180] In step 1402, user inputs are received to tag one or more objects shown in one or more portions of the 3D rendering displayed in step 1401. A given user input to tag an object can be performed according to step 1104.Attorney Docket No. FWIG-003W001

[0181] In optional step 1403, input data for one or more GUI element(s) for each tagged object are received. Input data for one or more GUI element(s) for a given tagged object can be performed according to optional step 1107.

[0182] In step 1404, data for the tagged object(s) are stored in non-volatile memory. The data stored for each tagged object can be the same as described in step 1108. The data for the tagged object(s) can be stored locally and can be synchronized with property features data in a property features data store, as discussed below.

[0183] In step 1405, a request is received for a virtual inspection report for the building structure. The request can be received in response to user input such as through a GUI element on the display.

[0184] In step 1406, virtual inspection report data are generated in response to the request received in step 1405. The virtual inspection report data represent a virtual inspection report. The virtual inspection report data can corresponding to at least some of the tagged objects in a property features data stored locally or in a property features data store.

[0185] In step 1407, the virtual inspection report is displayed on a display. In one or more embodiments, the virtual inspection report can comprise a table that can include a row for each tagged object. The table can include column(s) that represent respective GUI element(s) and the respective input data for each tagged object. The table can also include an image (e.g., image 1310) of the tagged object and / or an image (e.g., image 1320) of an annotated polygon or footprint of the building structure that represents the location of the tagged object with respect to the building structure. An example of a virtual inspection report 1500 that includes a table 1510 is shown in Fig. 14A. Additionally or alternatively, the virtual inspection report can include multiple sections with each section including tagged objects in the same category and / or with each section including tagged objects on or near the same outside wall of the building structure. The virtual inspection report can include an image (e.g., image 1310) of each tagged object and / or an image (e.g., image 1320) of an annotated polygon or footprint of the building structure that represents the location of each tagged object with respect to the building structure. An example of a virtual inspection report 1520 that includes one or more sections 1530 is shown in Fig. 14B. The virtual inspection report 1520 can be generated using property features data stored locally and / or in a property features data store, as discussed below.Attorney Docket No. FWIG-003W001

[0186] Fig. 15 is a flow chart of a computer-implemented method 1600 for detecting and / or segmenting structural features in digital images according to one or more embodiments. Method 1600 can be performed by one or more hardware-based processors in a computer for example by executing computer-readable instructions stored in the non-volatile memory of the computer.

[0187] In step 1601, digital images that collectively represent all outside walls of a building structure are received or provided. The digital images can be stored in and / or received from a local or remote computer memory (e.g., non-volatile memory). The digital images comprise digital image data that represent the digital images. The digital images can comprise or represent digital photographs or sampled digital images from one or more digital videos.

[0188] In step 1602, the digital images are fed into one or more trained models that is / are configured to segment and / or detect the structural features shown or represented in the digital images. The trained model(s) can comprise one or more trained computer vision segmentation model(s) and / or one or more separate trained machine-learning (ML) model(s). The computer vision segmentation model(s) and / or trained ML model(s) is / are trained with (a) digital images with known locations and shapes of structural features in example building structures and (b) digital images of example building structures that do not include any structural features. The computer vision segmentation model(s) and / or the trained ML model(s) can distinguish between each type of structural feature (e.g., between doors, windows, decks, balconies, vents, soffits, garage doors, and / or other type(s) of structural features).

[0189] In one or more embodiments, the structural features are segmented and / or detected using multiple (e.g., a plurality of) trained computer vision segmentation models and / or multiple (e.g., a plurality of) trained ML models. In one or more embodiments, a separate trained computer vision segmentation model and / or a separate trained ML model can be used to detect and / or segment each type of structural feature on / in a building structure. For example, a first trained computer vision segmentation model and / or a first trained ML model can be used to detect and / or segment windows on a building structure. Additionally or alternatively, a second trained computer vision segmentation model and / or a second trained ML model can be used to detect and / or segment doors on a building structure. Additionally or alternatively, a third trained computer vision segmentation model and / or a third trained ML model can be used to detect and / or segment soffits in a building structure. Each separate trained computer vision segmentation model and / or each separate trained ML model can beAttorney Docket No. FWIG-003W001 trained using (a) first sample images that include the type of feature and (b) second sample images that do not include the type of feature.

[0190] In step 1603, the structural features that are segmented and / or detected in each digital image can be masked (e.g., with pixels in a particular solid color). In one aspect, each different type of feature can be masked with a different color. For example, windows can be masked with a first color (e.g., red), and doors can be masked with a second color (e.g., blue). The masked digital images can be stored in computer memory (e.g., non-volatile memory).

[0191] In step 1604, a 3D rendering of the building structure is generated using the masked digital images. The 3D rendering of the building structure can be generated according to method 10. The masked color(s) of the structural features in the 3D rendering can visually highlight their respective locations during a virtual inspection. Image data for the 3D rendering of the outside walls of the building structure can be stored locally or in a property features data store, or in another location accessible through a unique link provided in the property features data.

[0192] Fig. 16 is an example 3D sparse point cloud rendering 1700 of a building structure 600 that includes one or more masked colors 1710 for the structural features 1720 according to one or more embodiments. The masked color(s) 1710 is / are represented as clusters of specifically colored points making it possible to find the 3D location for the structural features 1720 in space by scanning for dense regions of points with that specific color(s). In one or more embodiments, as noted above, each different type of structure feature 1720 can be represented in a different masked color 1710.

[0193] With reference now to Fig. 17, in one or more embodiments bounding boxes 1810 can be estimated or determined from the clusters of specifically colored points 1820 (e.g., red or another color) corresponding to structural features 1830 in a 3D sparse point cloud rendering 1800 of a building structure 600, for example as shown in Fig. 17. The bounding boxes 1810 can be used to automatically estimate the size and geospatial location of each structural feature 1830. The size and location of each structure feature 1830 can be stored in computer memory (e.g., non-volatile memory).

[0194] Fig. 18 is a simplified view of a display screen 1900 that displays a viewable portion 1940 of a three-dimensional rendering of a building structure 600 where each type of structural feature 620 is masked in a different color, shade and / or fill pattern (e.g., has a different state). For example, the windows 622 are masked in a first color (or have a firstAttorney Docket No. FWIG-003W001 state), the door 624 is masked in a second color (or has a second state), the vent 628 is masked in a third color (or has a third state), and the garage door 630 is masked in a fourth color (or has a fourth state). The viewable portion 1940 is the same as the viewable portion 840 except for the masked colors of the structural feature 620 shown in the viewable portion 1940.

[0195] Fig. 19 is a flow chart of a method 2000 for remotely monitoring a building structure for compliance with wildfire mitigation recommendations.

[0196] In step 2001, first digital images that collectively represent all outside walls of a building structure are received or provided. The first digital images represent the state of the building structure on a first time or a first date. The first digital images can be stored in and / or received from a local or remote computer memory (e.g., non-volatile memory). The first digital images comprise first digital image data that represent the first digital images. The first digital images can comprise or represent digital photographs or sampled digital images from one or more digital videos.

[0197] In step 2002, a first 3D rendering of the building structure is generated using the first digital images. The first 3D rendering of the building structure can be generated according to method 10 of Fig. 1.

[0198] In step 2003, a first virtual inspection of the building structure is performed using the first 3D rendering. The first virtual inspection can be performed according to method 70 of Fig. 6, method 1100 of Fig. 10, and / or method 1400 of Fig. 13.

[0199] Based at least in part on the first virtual inspection performed in step 2003, a virtual inspection report can be provided to an occupant, homeowner, and / or service provider to perform wildfire mitigation actions detailed in the virtual inspection report (e.g., in the remedial instructions).

[0200] In step 2004, second digital images that collectively represent all outside walls of the building structure are received or provided. The second digital images represent the state of the building structure on a second time or a second date that occurs after the first time / date (and presumably after at least some of the wildfire mitigation actions are completed). The second digital images can be stored in and / or received from a local or remote computer memory (e.g., non-volatile memory). The second digital images comprise second digital image data that represent the second digital images. The second digital first images canAttorney Docket No. FWIG-003W001 comprise or represent digital photographs or sampled digital images from one or more digital videos.

[0201] In step 2005, a second 3D rendering of the building structure is generated using the second digital images. The second 3D rendering of the building structure can be generated according to method 10 of Fig. 1.

[0202] In step 2006, a second virtual inspection of the building structure is performed using the second 3D rendering. The second virtual inspection can be performed according to method 70, method 1100, and / or method 1400. The second virtual inspection can be performed to confirm (e.g., remotely confirm) whether or not respective wildfire mitigation actions have been performed (e.g., pursuant to the virtual inspection report following the first virtual inspection).

[0203] In step 2007, it is determined whether any wildfire mitigation actions still need to be performed. For example, there may be one or more wildfire mitigation actions that have not been performed and / or additional wildfire mitigation actions identified during the second virtual inspection. If there are no wildfire mitigation actions that still need to be performed for the building structure (i.e., step 2007=no), then in step 2008 it is determined that the building structure is in compliance with wildfire mitigation recommendations (e.g., provided in a virtual inspection report). If there is / are one or more wildfire mitigation actions that still need to be performed for the building structure (i.e., step 2007=yes), then the method 2000 can return to step 2004 in a loop 2008 to receive new digital images that collectively represent all outside walls of the building structure. The new digital images can be received preferably after any remaining wildfire mitigation action(s) for the building structure are completed. An additional virtual inspection report can be provided to an occupant, homeowner, and / or service provider to confirm whether or not any remaining wildfire mitigation action(s) (e.g., as detailed in the remedial instructions of one or more previous virtual inspection reports) have been sufficiently addressed.

[0204] Fig. 20 is a flow chart of a method 2100 for performing a virtual fire inspection according to one or more embodiments.

[0205] In step 2101, a geospatially calibrated overhead or aerial digital image of a property including a building structure is acquired or received. The overhead / aerial image can be a satellite image, an aerial image (e.g., captured by plane, drone, or balloon) or another aerial images of the property including the building structure. The aerial images can be high-qualityAttorney Docket No. FWIG-003W001 and / or can be scaled to determine the geospatial location and dimensions of the structure and objects on the property such as trees and / or shrubs and the respective distance between the nearest outside wall of the property and each object (e.g., each tree and / or shrub).

[0206] Fig. 21 shows an example geospatially calibrated overhead image 2200 of a property 2210 including a building structure 2220. The geospatially calibrated overhead image 2200 can be a satellite image or another image. The overhead image 2200 shows large vegetation such as trees 2230 and shrubs 2240. In one or more embodiments, the overhead image 2200 can show one or more detached (e.g., ancillary) structures 2225 on the property 2210. The detached structure(s) 2225 can include a detached garage, a shed, a pool house, and / or another detached structure. The dimensions of the overhead image 2200 overall, as well as respective sizes and dimensions of the objects represented in the overhead image 2200 (such as the building structure 2220, the trees 2230, the shrubs 2240, and any ancillary structure(s) 2225) can be measured and / or determined as described, for example, in U.S. Application Publication No. 2023 / 0023808, previously incorporated by reference. In addition, the respective distances between the building structure 2220 and other objects (e.g. , each tree 2230, each shrub 2240, and each ancillary structure 2225) can be measured and / or determined as described, for example, in U.S. Application Publication No. 2023 / 0023808, previously incorporated by reference.

[0207] Returning to Fig 20, in optional step 2102 a geospatial perimeter of the building structure can be determined using the geospatially calibrated aerial digital image. The geospatial perimeter of the building structure can be determined using a trained ML structure footprint detection model that is configured to detect the structure footprint in an overhead image of a building structure (for example as described in step 104 of Fig. 1).

[0208] In step 2103, fuel sources represented in the aerial image are detected. The fuel sources can be detected with one or more trained ML models that is / are configured to detect fuel sources, such as trees, shrubs, and / or detached structures in aerial images.

[0209] In step 2104, the property ignition risks, due to a wildfire, for the building structure are determined or estimated. The property ignition risks can be determined or estimated using a PIM, for example as disclosed in U.S. Application Publication No. 2023 / 0023808, previously incorporated by reference.

[0210] In step 2105, digital images that collectively represent all outside walls of a building structure are received or provided. The digital images can be stored in and / or received from aAttorney Docket No. FWIG-003W001 local or remote computer memory (e.g., non-volatile memory). The digital images comprise digital image data that represent the digital images. The digital images can comprise or represent digital photographs or sampled digital images from one or more digital videos.

[0211] In step 2106, a 3D rendering of the building structure is generated using the digital images received in step 2105. The first 3D rendering of the building structure can be generated according to method 10 of Fig. 1.

[0212] In step 2107, a virtual inspection of the building structure is performed using the 3D rendering. The virtual inspection can be performed according to method 70, method 1100, and / or method 1400. The virtual inspection can identify one or more wildfire mitigation actions to be performed as recommendations for compliance with wildfire insurance policy requirement; for example, the wildfire mitigation action(s) can represent wildfire insurance policy violations and / or exclusions. The virtual inspection can be performed using a virtual wildfire inspection apparatus 3900 of Fig. 39, discussed below.

[0213] In step 2108, the property ignition risks, due to a wildfire, for the building structure are updated (e.g., determined a second time) based on the assumption or expectation that the wildfire mitigation action(s) are performed.

[0214] In step 2109, a virtual inspection report can be generated. The virtual inspection report can be shared, displayed, and / or printed. The virtual inspection report can be generated can be generated using property features data stored locally and / or in a property features data store, as discussed below.

[0215] The virtual inspection report and / or a 3D rendering of the building structure that includes tagged objects can be provided to a person (via the person’s device) before and / or when the person arrives at the property to perform wildfire mitigation actions. The person performing the wildfire mitigation actions can be a service provider or a homeowner.

[0216] Fig. 22 is a flow chart of a computer-implemented method 2300 for guiding a person such as a service provider, via a mobile computing device carried by or associated with the person, to perform wildfire mitigation actions for a building structure according to one or more embodiments.

[0217] In step 2301, a 3D rendering of the outside walls of the building structure can be displayed on a mobile computing device carried by and / or associated with the service provider. The service provider can be a third-party service provider, a homeowner, or another person. The 3D rendering can comprise or can be a volumetrically-rendered reconstruction ofAttorney Docket No. FWIG-003W001 the outside walls of a building structure such as a volume-rendered reconstruction 60 as described herein. In one or more embodiments, only a portion of the 3D rendering of the building structure can be displayed, for example according to a virtual perspective and virtual position of a virtual service provider with respect to the 3D rendering of the building structure. The portion of the 3D rendering of the building structure can be displayed as described with respect to Figs. 7A and 7B where reference to the virtual inspector can be replaced with reference to a virtual service provider. Additionally, or alternatively, the displayed portion of the 3D rendering of the building structure can be updated in response to user inputs, for example as described in method 70 (Fig. 6) and Figs. 7A, 7B, 8A, 8B, 9A, and 9B where reference to the virtual inspector can be replaced with reference to a virtual service provider.

[0218] The 3D rendering of the building structure can include a 3D rendering and / or a 2D rendering (e.g., an aerial view) of the property surrounding the building structure and any objects on the property. The property and objects shown in the 2D and / or 3D rendering correspond to the property and objects captured in the digital images that were used as inputs to generate the 3D rendering of the building structure, for example as described in method 10 of Fig. 1. Image data for the 3D rendering of the outside walls of the building structure can be included in property features data stored locally or in a property features data store, or in a unique link provided in the property features data.

[0219] In step 2302, one or more tagged objects can be displayed in the 3D rendering displayed according to step 2301. The tagged object(s) represent physical objects and / or physical locations (e.g., areas) where one or more wildfire mitigation actions should be performed. The tagged object(s) can be tagged previously during a virtual inspection that was performed, for example, according to method 70 of Fig. 6, method 1100 of Fig. 10, and / or method 1400 of Fig. 13. The tagged object(s) can be displayed according to the portion of the 3D rendering of the building structure displayed on the mobile computing device carried by the service provider (e.g., according to the virtual perspective and virtual position of a virtual service provider with respect to the 3D rendering of the building structure). In one or more embodiments, only the tagged object(s) viewable in the displayed portion of the 3D rendering of the building structure is / are displayed. In one or more other embodiments, one or more (e.g., some or all) of the tagged objects not viewable in the displayed portion of the 3D rendering of the building structure can be displayed in addition to the tagged object(s) viewable in the displayed portion of the 3D rendering of the building structure. DataAttorney Docket No. FWIG-003W001 corresponding to the indexed position of each tagged object can be provided from property features data stored in a property features data store, which can be stored locally, as described herein.

[0220] The tagged object(s) can be visually highlighted in the displayed portion of the 3D rendering. For example, the color, shading, and / or fill pattern (or other displayed visible attribute) of a tagged object 1200 can be different than that of an untagged object, for example as described with respect to step 1105 of method 1100 of Fig. 10 and / or as shown in Fig. H.

[0221] In step 2303, a list of the tagged objects can be displayed on the mobile computing device carried by the service provider. The list can include a description of the physical object associated with each tagged object. The list can include or can consist of the tagged object(s) that is / are shown in the displayed portion of the 3D rendering of the building structure (e.g., displayed in steps 2301 and 2302). Alternatively, the list can include all tagged object(s).

[0222] The list can include a link for each tagged object that, when activated: (a) updates the displayed portion of the 3D rendering so as to visually show the tagged object associated with the activated link and / or (b) displays additional details associated with the tagged object. The additional details associated with the tagged object can include a description of the physical object associated with the tagged object, the physical location corresponding to the tagged object on the physical property and / or on the physical building structure, a virtual location of the tagged object with respect to the 3D rendering of the building structure, and / or the priority for the tagged object. The list can be or comprise a partial list of the tagged objects in one or more embodiments. The list can be in the form of a table, such as the table 1510 shown in Fig. 14A. The list or table can include additional or fewer columns than the table 1510. Data describing each tagged object, such as the data shown in table 1510, can be provided from (and stored in) a property features data store, which can be stored locally, as described herein.

[0223] In one or more embodiments, the list or partial list of tagged objects can be included in a virtual inspection report that can be generated previously or on the fly (e.g., in response to user input or a request) for example according to method 1400 of Fig. 13 (e.g., step 1406) and / or as shown in Fig. 14A and / or in Fig. 14B. Additionally or alternatively, activation of a link associated with a tagged object in the list or partial list can cause a corresponding portionAttorney Docket No. FWIG-003W001 of a virtual inspection report to be displayed, the corresponding portion including the tagged object.

[0224] In one or more embodiments, the list or partial list of tagged objects can be displayed concurrently with the displayed portion of the 3D rendering (e.g., displayed in step 2301) and / or with the tagged object(s) displayed in the 3D rendering (e.g., displayed in step 2302). The partial list can correspond to only the tagged object(s) displayed in the 3D rendering in one or more embodiments.

[0225] Fig. 23 shows an example viewable portion 2440 of a building structure 600, as shown on a display 80, that includes several tagged objects 2400. Each tagged object 2400 can be the same as a tagged object 1200 as shown in Fig. 11. For example, the visual appearance of each tagged object 2400 can be different than the visual appearance of other objects (untagged objects) shown in the viewable portion 2440 of a three-dimensional rendering of a building structure 600 such as the window 622, door 624, and outside wall 610A. In one or more embodiments, a list or partial list 2410 of tagged objects can be displayed simultaneously with the viewable portion 2440 of the building structure 600.Additionally or alternatively, the list or partial list 2410 can be displayed only when (e.g., in response to) user input that selects one of the tagged objects (e.g., in response to receiving a selection of a tagged object in step 2304 of Fig. 22, discussed below).

[0226] Returning to Fig. 22, in step 2304, a selection of a tagged object is received. A tagged object can be selected by user input received by the mobile computing device carried by or associated with the service provider. The user input can include an audible user input (e.g., a voice command), a touchscreen input (e.g., a touch or touch gesture), a keyboard input (e.g., from a virtual keyboard displayed on the mobile computing device or from a physical keyboard), a mouse input, and / or another user input.

[0227] Alternatively, a tagged object can be selected automatically. Automatic selection of a tagged object can be performed based, at least in part, on the current geospatial position of the service provider (as indicated by the geospatial position of the mobile computing device carried by or associated with the service provider), the geospatial positions associated with the tagged objects, and / or the priorities of the tagged objects. In one example, respective physical distances can be calculated between the current geospatial position of the service provider (as indicated by the geospatial position of the mobile computing device carried by or associated with the service provider) and the respective geospatial position associated withAttorney Docket No. FWIG-003W001 each tagged object. The tagged object that, in the physical world, is physically closest to the current geospatial position of the service provider (as indicated by the geospatial position of the mobile computing device carried by or associated with the service provider) can be automatically selected. In another example, the automatic selection of a tagged object can be performed according to the priorities associated with the tagged object where the highest (most urgent) priority tagged object is automatically selected. If multiple tagged objects have the same priority, the tagged object that, in the physical world, is physically closest to the current geospatial position of the service provider (as indicated by the geospatial position of the mobile computing device carried by or associated with the service provider), relative to the respective physical distances from the current geospatial position of the service provider to the other tagged object(s) having the same priority, can be automatically selected.

[0228] In one or more embodiments, the visual appearance of a tagged object can be changed when the tagged object is selected. Fig. 24 shows an example viewable portion 2540 of a three-dimensional rendering of a building structure 600. The viewable portion 2540 in Fig. 24 is the same as the viewable portion 2440 in Fig. 23 except that the viewable portion 2540 in Fig. 24 includes a selected tagged object 2500. The visual appearance of the selected tagged object 2500 is different than the visual appearance of the tagged objects 2400 that have not been selected. The altered / updated visual appearance of the selected tagged object 2500 can be referred to as a selected state of the selected tagged object 2500.

[0229] Returning to Fig. 22, in step 2305, the physical (e.g., real-world) location associated with the selected tagged object is provided. The physical location can be provided as an image representing a virtual location of the selected tagged object relative to the 3D rendering of the building structure, such as in an overhead image / overhead rendering 1320 in Fig. 12. Additionally or alternatively, the physical location can be provided as a description of the physical location associated with the selected tagged object (e.g., “north-facing side of building structure,” “next to front door,” etc.). Additionally or alternatively, the physical location can be provided as a geospatial position (e.g., geospatial position data) associated with the selected tagged object.

[0230] In one or more embodiments, the system can compare the current geospatial position of the service provider (as indicated by the geospatial position of the mobile computing device carried by or associated with the service provider) and the geospatial position associated with the selected tagged object and provide directions to direct the service provider to the real -world location (e.g., geospatial location) for the selected tagged object.Attorney Docket No. FWIG-003W001 The directions can be provided using graphics, text, audio (e.g., voice), and / or another form. The current geospatial position of the service provider and the geospatial position of the selected tagged object can be iteratively compared to iteratively update the directions as the service provider and the associated mobile computing device move towards the real-world location for the selected tagged object.

[0231] In step 2306, an image of the selected tagged object can be displayed. The displayed image can comprise a detailed or zoomed-in image of selected tagged object. The displayed image can comprise a virtual image from the 3D rendering (e.g., captured during virtual inspection) such as an image 1310 in Fig. 12. Additionally or alternatively, the displayed image can comprise at least a portion of a digital image (e.g., used to generate the 3D rendering) that depicts the physical object (or area) that corresponds to the selected tagged object.

[0232] In step 2307, one or more wildfire mitigation action(s) associated with the tagged object is / are provided. The wildfire mitigation action(s) can comprise or can be emergency mitigation action(s), as discussed herein. Additionally or alternatively, the wildfire mitigation action(s) can comprise remedial instructions. In general, the wildfire mitigation action(s) represent physical action(s) to be performed by the service provider, with respect to the physical object (or area) that corresponds to the selected tagged object, to reduce the ignition risk of a building structure in the event of a wildfire. The wildfire mitigation action(s) was / were previously determined during a wildfire inspection of the building structure, such as during a virtual wildfire inspection of the building structure that was performed for example, according to method 70 of Fig. 6, method 1100 of Fig. 10, and / or method 1400 of Fig. 13. Examples of wildfire mitigation action(s) can include applying a fire retardant to an object or area (e.g., to a mulch bed and / or to vegetation such as a bush), sealing garage doors, cleaning gutters, moving outdoor furniture and / or grills, trimming or removing large vegetation such as trees, moving woodpiles away from the building structure, sealing vents, and / or other wildfire mitigation actions.

[0233] In step 2308, the wildfire mitigation action(s) can be tracked. In one or more embodiments, the wildfire mitigation action(s) can be tracked by monitoring or tracking the geospatial position of the service provider (as indicated by the geospatial position of the mobile computing device carried by or associated with the service provider) and comparing the geospatial position of the service provider to the geospatial position associated with the selected tagged object. The geospatial position of the service provider can be used as a log ofAttorney Docket No. FWIG-003W001 the service provider’s actions to provide an indication that the wildfire mitigation action(s) (or at least some of the wildfire mitigation action(s)) was / were performed on / in the physical object (or area) associated with the selected tagged object.

[0234] For example, when the geospatial position of the service provider is within (e.g., less than or equal to) a predetermined physical distance of the geospatial position associated with the selected tagged object, it can be assumed that service provider performed the wildfire mitigation action(s) (or at least some of the wildfire mitigation action(s)) on / in the physical object (or area) associated with the selected tagged object. When the geospatial position of the service provider is not within (e.g., greater than) the predetermined physical distance of the geospatial position associated with the selected tagged object, it can be assumed that the service provider did not perform the wildfire mitigation action(s) (or at least some of the wildfire mitigation action(s)) on / in the physical object (or area) associated with the selected tagged object.

[0235] In one or more examples, a geofence can be established around or with respect to the physical object (or area) associated with the selected tagged object. When the service provider enters a geofenced region, as indicated by the geospatial position of the service provider, it can be assumed that service provider performed the wildfire mitigation action(s) (or at least some of the wildfire mitigation action(s)) on / in the physical object (or area) associated with the selected tagged object. When the service provider does not enter the geofenced region, as indicated by the geospatial position of the service provider, it can be assumed that service provider did not perform the wildfire mitigation action(s) (or at least some of the wildfire mitigation action(s)) on / in the physical object (or area) associated with the selected tagged object.

[0236] Additionally or alternatively, the wildfire mitigation action(s) can be tracked by prompting the service provider (via the mobile computing device carried by or associated with the service provider) to capture one or more digital images (e.g., one or more digital photographs and / or one or more digital videos) of the physical object (or area) associated with the selected tagged object after the wildfire mitigation action(s) is / are performed. In one or more examples, the prompt can be provided after or in response to an indication that the service provider has moved to the physical object (or area) associated with the selected tagged object, for example by comparing the geospatial position of the service provider and the geospatial position of the physical object (or area) associated with the selected tagged object or by comparing the geospatial position of the service provider and a geofence for theAttorney Docket No. FWIG-003W001 physical object (or area) associated with the selected tagged object. The prompt may also be provided after or in response to receiving a manual input (e.g., a manual confirmation) from the mobile computing device carried by or associated with the service provider, the manual input indicating that the service provider has moved to the physical object (or area) associated with the selected tagged object and / or that the service provider has performed the wildfire mitigation action(s).

[0237] In one or more embodiments, a prompt can be provided, via the mobile computing device carried by or associated with the service provider, that requests, in step 2309, that the service provider confirm that the wildfire mitigation action(s) was / were performed. In step 2310, the service provider can confirm, via the mobile computing device, that the wildfire mitigation action(s) was / were performed. Confirmation can be provided by user input on the mobile computing device, such as from a user input device (e.g., touch screen, mouse, keyboard, microphone (voice command), etc.) in communication with the mobile computing device.

[0238] In one or more embodiments, the prompt or request, provided in step 2309, for confirmation that the wildfire mitigation action(s) was / were performed can include a request for the service provider to capture one or more digital images of the physical object (or area) to show the completion of the wildfire mitigation action(s) associated with the selected tag.

[0239] After it appears that the wildfire mitigation action(s) was / were performed during wildfire mitigation action tracking in step 2308 and / or the service provider confirmed that the wildfire mitigation action(s) was / were performed in step 2310, the status of the selected tagged object can be updated in optional step 2311 (via placeholder A). The status of the selected tagged object can include a visual appearance and / or a virtual status.

[0240] In one or more embodiments, the visual appearance of a tagged object, as it appears on the display screen, can have a first state that indicates that wildfire mitigation action(s) need to be performed and second state that indicates that the wildfire mitigation action(s) has / have been performed. The first state can alternately be referred to as an initial state or an incomplete state. The second state can alternately be referred to as a completed state. The visual appearance of all tagged objects are in the first state when the service provider arrives at the property to begin work. As the service provider moves about the property and it appears that the wildfire mitigation action(s) was / were performed during wildfire mitigation action tracking for a selected tagged object in step 2308 and / or the service providerAttorney Docket No. FWIG-003W001 confirmed that the wildfire mitigation action(s) for the selected tagged object was / were performed, in response to a confirmation request in step 2309, the visual appearance of the selected tagged object transitions from the first state to the second state.

[0241] The visual appearance of each tagged object shown in the 3D rendering, displayed in steps 2301 and 2302, can transition between the first and second states.

[0242] In one or more embodiments, the visual appearance of each tagged object can transition between an initial / incomplete state (e.g., the first state discussed above), a selected state (e.g., as discussed above with respect to Fig. 24), and a completed state (e.g., the second state discussed above). Fig. 25 shows an example viewable portion 2640 of a building structure 600. The viewable portion 2640 in Fig. 25 is the same as the viewable portion 2540 in Fig. 24 except that the viewable portion 2640 in Fig. 25 includes a completed tagged object 2600 instead of a selected tagged object 2500. The visual appearance (e.g., a completed state) of the completed tagged object 2600 in Fig. 25 is different than the visual appearance (e.g., a selected state) of the selected tagged object 2500 in Fig. 24, the visual appearance (e.g., an initial / incomplete state) of the tagged objects 2400 that have not been selected, and the visual appearance (e.g., an untagged state) of other objects (untagged objects) shown in the viewable portion 2440 of the building structure 600 such as the window 622, door 624, and outside wall 610A.

[0243] Additionally or alternatively, the visual appearance of each tagged object in the list of tagged objects, displayed in step 2303, can transition between the first and second states. The first and second states can be the same or different for tagged objects shown in the 3D rendering than for tagged object in the list of tagged objects. Additionally or alternatively, the visual appearance of the tagged objects in the list can transition between an initial / incomplete state, a selected state, and a completed state.

[0244] Additionally or alternatively, the virtual status of each selected tagged object can be updated, such as in a field in a database and / or in a table.

[0245] In step 2312, the system determines whether there are any additional tagged objects to be selected. If so (i.e., step 2312=yes), the method 2300 can return to step 2304 (via placeholder B) in a loop where another tagged object is selected in the next iteration of step 2304. In one or more embodiments, the method 2300 can return to optional step(s) 2301, 2302, and / or 2303 before returning to step 2304 from step 2312. For example, if the system determines there are one or more additional tagged objects to be selected in step 2312, theAttorney Docket No. FWIG-003W001 method 2300 can return to optional steps 2301 and 2302 where a displayed portion of the 3D rendering of the building structure including one or more tagged objects is displayed on the mobile computing device carried by and / or associated with the service provider. In the displayed portion of the 3D rendering, the visual appearance of any tagged objects that was / were previously selected and for which the mitigation action(s) has / have already been completed can be different from the visual appearance of any tagged objects that has / have not yet been selected. The next tagged object can be selected in step 2304 directly from the displayed portion of the 3D rendering of the building structure (displayed in optional steps 2301 and 2302) or from a list of tagged objects that can be displayed in optional step 2303.

[0246] The loop from step 2312 to optional step(s) 2301, 2302, and / or 2303 and step 2304 can be repeated iteratively until all tagged objects have been selected and the respective mitigation action(s) has / have been performed. When the system determines that there are no additional tagged objects to be selected (i.e., step 2312=no), the method 2300 can proceed to optional step 2313 to request the service provider, via one or more prompts on the mobile computing device carried by and / or associated with the service provider, to capture digital images that collectively represent all outside walls of the building structure including the physical locations associated with the selected tagged objects. The prompt(s) can instruct the user to: 1) capture the entire vertical span of the building structure during image acquisition; 2) avoid quick movements or panning with the device during image acquisition; 3) make at least one full loop around the perimeter of the building structure while acquiring the digital images; 4) capture the property surrounding the building structure during image acquisition; 5) stand a predetermined distance (e.g., 30-50 feet) from the building structure to capture the nearfield property during image acquisition; and / or 6) perform a 360-degree pan of the digital camera (or digital video camera) at least once on each side of the building structure. In some embodiments, these constraints are validated during image acquisition, and the system provides real-time feedback to the user when constraints are violated.

[0247] Alternatively, control signals can be sent to a drone or robot to capture the digital images in which case the step 2313 may be performed by sending the control signals to the drone or robot.

[0248] The digital images captured by the mobile computing device, a drone, or a robot can be used to generate a 3D rendering of the building structure and surrounding property. This 3D rendering of the building structure and surrounding property can be used to perform a virtual inspection to confirm that all mitigation actions have been performed, for exampleAttorney Docket No. FWIG-003W001 according to method 2000 of Fig. 19 where the digital images captured by the mobile computing device, a drone, or a robot can represent a second time or a second date in steps 2004-2008, as described herein.

[0249] In step 2314, the digital images (requested in step 2313) that collectively represent all outside walls of the building structure including the physical locations associated with the selected tagged object are received. In one or more embodiments, the digital images can be received by locally storing the digital images on the mobile computing device carried by or associated with the service provider. Additionally or alternatively, the digital images can be received by sending the digital images to a computer (e.g., a server or another computer) in communication (e.g., in wired or wireless communication) with the mobile computing device carried by or associated with the service provider.

[0250] Fig. 26 is a block diagram of a system 2700 for performing a virtual inspection of a building structure according to one or more embodiments. The system 2700 includes a first data store 2701, an image processing engine 2702, a second data store 2703, an annotation engine 2704, and a wildfire risk modelling engine 2705.

[0251] The first data store 2701 is configured to receive and store image data received from one or more image sources 2610. The image source(s) 2610 can include a ground-level image source and an overhead image source. The first data store 2701 can comprise non-volatile computer memory that can be located in a computer or a server or distributed across one or more computers and / or servers.

[0252] The ground-level source can include a mobile computing device having a camera that captured digital images (e.g., digital photos and / or digital video images) that collectively represent all outside walls of a target building structure to be virtually inspected. The digital images from the ground-level source can further include at least a portion of the target property on which the building structure is located. For example, the mobile computing device can be positioned at a distance such that at least some of the property is viewable in the foreground of the digital images during image capture. In one or more embodiments, the property up to at least a predetermined distance from the outside walls of the building structure can be captured, such as in a range of 5 feet to 50 feet from the building structure (including any range or value therebetween) can be captured in the digital images. The captured portion of the property can correspond to a zone in which fuel sources located within the zone can cause an increased risk of ignition for the building structure due to aAttorney Docket No. FWIG-003W001 wildfire. It is noted that the digital images collected from the ground-level source can be acquired above ground level, such as from a drone, provided that the digital images collectively include the outside walls of the building structure and optionally at least a portion of the property on which the building structure is located.

[0253] The overhead image source can include a repository or database of geospatially calibrated overhead images including a geospatially calibrated overhead image of the target property including the target building structure. The geospatially calibrated overhead images can be acquired from a camera on / in a satellite, an airplane, a drone, or a balloon. The geospatially calibrated overhead images include geospatial data such that an image location in a geospatially calibrated overhead image has a corresponding geospatial position or location. Additionally or alternatively, the overhead image source can include a repository or database of geospatially calibrated polygons representing geospatial perimeters of building structures including a target geospatial perimeter of the target building structure. Geospatially calibrated polygons can also define perimeters of trees and / or other major vegetation on the properties.

[0254] The image processing engine 2702 is in communication (e.g., network or local) communication with the first data store 2701. The image processing engine 2702 includes one or more image processors that is / are configured to process the image data 2720 received from the first data store 2701. The one or more image processors can detect the outside walls of the building structure and structural features (e.g., structural features 620 such as windows 622, doors 624, decks 626, balconies, vents 628 (e.g., air vents, laundry vents, and / or other vents), soffits, and / or garage doors 630 as shown in Fig. 6) represented in the image data 2720 from the ground-level source. For example, the image processor(s) can detect the outside walls of the building structure and structural features by feeding the digital images (e.g., image data 2720) into one or more trained ML models running on the image processor(s). The trained ML model(s) can determine or estimate the type or category of each structural feature, the materials of each structural feature, the materials of each outside wall, and / or respective indexed image locations (e.g., indexed structural locations) of the outside walls and the structural features. The image processing engine 2702 can generate encoded structural descriptors that represent the type or category of each structural feature, the materials of each structural feature, the materials of each outside wall, and / or respective indexed image locations (e.g., indexed structural locations) of the outside walls and the structural features. The encoded structural descriptors can be stored in a second data storeAttorney Docket No. FWIG-003W001 2703. The second data store 2703 can comprise non-volatile computer memory that can be located in a computer or a server or distributed across one or more computers and / or servers.

[0255] The one or more image processors can also detect one or more fuel sources that is / are located in the at least a portion of the target property on which the building structure is located and that is / are represented in the image data 2720 from the ground-level source. For example, the image processor(s) can detect the fuel source(s) by feeding the digital images (e.g., image data 2720) into one or more trained ML models running on the image processor(s). The trained ML model(s) used to detect the fuel sources can be the same or different than the trained ML model(s) used to detect the outside walls and / or the structural features. The trained ML model(s) can determine or estimate the type or category of each fuel source, the materials of each fuel source, the size of each fuel source, and / or respective indexed image location(s) (e.g., indexed fuel location(s)) of the fuel source(s)). The image processing engine 2702 can generate encoded fuel descriptors that represent the type or category of each fuel source, the materials of each fuel source, and / or respective indexed image location(s) (e.g., indexed fuel location(s)) of the fuel source(s). Examples of fuel sources include trees, large vegetation, wooden furniture, wooden fences, secondary structures (e.g., sheds, detached garages), debris such as leaves, leaf piles, and / or wood piles, and / or fuel tanks (e.g., propane tanks). The encoded fuel descriptors can be stored in the second data store 2703.

[0256] In one or more embodiments, the one or more image processors can also determine a geospatially calibrated perimeter (e.g., a geospatially calibrated polygon 420) of the target building structure using the aerial image(s), for example as described in step 2102 of Fig. 20 and / or in step 104 of Fig. 1. Additionally or alternatively, a geospatially calibrated perimeter can be provided by and / or received from a third party. One or more encoded geospatial descriptors of the geospatial perimeter of the target building structure can be stored in the second data store 2703.

[0257] The image processing engine 2702 can receive as inputs the digital images that collectively represent all outside walls of the target building structure and the geospatial perimeter of the target building structure and can generate a geospatially aligned three-dimensional rendering of the target building structure. The geospatially aligned three-dimensional rendering of the target building structure can be generated according to method 10 of Fig. 1. The geospatially aligned three-dimensional rendering of the target building structure can be stored in the second data store 2703.Attorney Docket No. FWIG-003W001

[0258] The annotation engine 2704 is configured to display, on a display screen 2706 in communication with the annotation engine 2704, a portion of the geospatially aligned three-dimensional rendering of the target building structure. The portion of the geospatially aligned three-dimensional rendering of the target building structure displayed on the display screen can be updated (e.g., iteratively updated) in response to user input, for example as a virtual inspector is moved relative to the three-dimensional rendering, for example as described in step 1401 of Fig. 13. One or more objects can be tagged, measured, and / or annotated during virtual inspection which can result in updates and / or additions to the encoded fuel descriptors and / or the encoded structural descriptors stored in the second data store 2703.

[0259] Fig. 27 shows an example viewable portion 2840 of a three-dimensional rendering of a building structure 600. The viewable portion 2840 is the same as the viewable portion 1040 of Fig. 9B except that in Fig. 27 the window 622 has been tagged (e.g., in response to user input or automatically) as a tagged object 2800. In one or more embodiments, the tagging can include generating a bounding region (e.g., a bounding box) corresponding to a perimeter 2822 of the window 622. When the window 622 is tagged manually, the window 622 can be, prior to tagging, an uncategorized object 2801.

[0260] The annotation engine 2704 (Fig. 26) is configured to receive one or more annotations 2850 in respective annotation fields for the tagged object 2800. A first annotation 2850A can include an indexed location of the tagged object 2800 including an indexed image location in the three-dimensional rendering of the building structure 600 and a corresponding indexed geospatial location for the tagged object 2800 (e.g., a real-world or physical location associated with the tagged object 2800). The indexed image location can include an image position and an image span that defines an image size of the tagged object 2800. The indexed geospatial location can include a geospatial position that defines a physical location and a physical size of the tagged object 2800 in the physical world.

[0261] The first annotation 2850A can be generated automatically according to the bounding region 2810 defined by user input(s). The bounding region 2810 and / or the first annotation 2850A can be automatically generated in one or more embodiments, for example, when the window 622 was automatically detected by the image processing engine 2702 (e.g., using one or more ML models). When the bounding region 2810 is automatically generated, the position and / or size of the bounding region 2810 can be manually adjusted (e.g., in response to user input) to further align the bounding region 2810 and the perimeter 2822 of the window 622. The manual adjustment automatically produces a corresponding adjustment ofAttorney Docket No. FWIG-003W001 the indexed location (e.g., the indexed image location and the indexed geospatial location) and respective size of the tagged object 2800.

[0262] In one or more embodiments, the annotation engine 2704 can receive one or more user inputs to provide additional annotations 2850 (e.g., 2850B . . . 2850N) to provide or define one or more additional properties of the tagged object 2800. For example, the annotations can include one or more categories for the tagged object 2800 (e.g., an outside wall, a structural feature, a type of structural feature, a fuel source, a type of fuel source), one or more materials of the tagged object 2800, and / or a location (e.g., floor of the building structure 600 or portion of the property) where the tagged object 2800 is located in the physical world. The category(ies), material(s), and / or location provided in the annotations 2850 can selected from predefined lists (e.g., check boxes, drop-down lists, etc.) and / or can be entered manually.

[0263] After the annotations 2850 are entered or provided, the annotations 2850 are provided as updates to the encoded structural descriptors stored in the second data store 2703. After the encoded structural descriptors are updated in the second data store 2703, the wildfire ignition risk modelling engine 2705 can model an ignition risk of the building structure due to wildfire using the encoded structural descriptors and the encoded fuel descriptors stored in the second data store 2703 as inputs to a wildfire ignition risk model that can be executed, using computer-readable instructions, by one or more risk processors in the wildfire risk modelling engine 2705. An example of a wildfire ignition risk model is a PIM. An output of the wildfire ignition risk modelling engine 2705 can be provided to and / or displayed on the display screen 2706.

[0264] Fig. 28 shows an example viewable portion 2840 of a three-dimensional rendering of a building structure 600. The viewable portion 2840 is the same as the viewable portion 2840 of Fig. 27 except that in Fig. 28 the tagged object 2800 is a wood pile 660 instead of a window 622. The bounding region 2810 corresponds to a perimeter 2822 of the wood pile 660. When the wood pile 660 is tagged manually, the wood pile 660 can be, prior to tagging, an uncategorized object 2801.

[0265] The annotation engine 2704 (Fig. 26) is configured to generate one or more annotations 2850 for the tagged object 2800 in the same or similar manner as discussed above though the entries, fields, or menu options for a fuel source such as a wood pile 660 may differ in one more respects from those for a structural feature 600 such as a window 622Attorney Docket No. FWIG-003W001

[0266] After the annotations 2850 are entered or provided for the tagged object 2800 (wood pile 660), the annotations 2850 are provided as updates to the encoded fuel descriptors stored in the second data store 2703. After the encoded fuel descriptors are updated in the second data store 2703, the wildfire ignition risk modelling engine 2705 can model an ignition risk of the building structure due to wildfire using the encoded structural descriptors and the encoded fuel descriptors stored in the second data store 2703 as inputs to the wildfire ignition risk model. An output of the wildfire ignition risk modelling engine 2705 can be provided to and / or displayed on the display screen 2706.

[0267] Fig. 29 is a block diagram of the image processing engine 2702 of Fig. 26 according to one or more embodiments. The image processing engine 2702 can include one or image detectors 2900, one or more encoders 2910, and a 3D image generator 2920. The one or more image detectors 2900 is / are configured to receive the image data 2620 and detect objects represented in the image data 2620, such as outside walls, structural features, and fuel sources. The objects can be detected using one or more trained ML models in one or more embodiments. The one or more encoders 2910 is / are configured to generate encoded descriptors that represent properties and / or indexed locations of the objects detected by the one or more image detectors 2900. The output of the one or more image detectors 2900 is provided as an input to the one or more encoders 2910.

[0268] In one or more embodiments, the one or more image detectors 2900 includes a structural feature detector 2901 and a fuel source detector 2902. The structural feature detector 2901 is configured to detect the outside walls and structural features on a building structure as represented in the image data 2620. The fuel source detector 2902 is configured to detect fuel sources and other violations on the property as represented in the image data 2620. The structural feature detector 2901 and the fuel source detector 2902 can each include one or more respective trained ML models.

[0269] In one or more embodiments, the one or more encoders 2910 includes a structural feature encoder 2911 and a fuel source encoder 2912. The input of the structural feature encoder 2911 is connected to the output of the structural feature detector 2901. The input of the fuel source encoder 2912 is connected to the output of the fuel source detector 2902. The structural feature encoder 2911 and the fuel source encoder 2912 can each include one or more respective trained ML models. The output of the one or more encoders 2910 (e.g., the outputs of the structural feature encoder 2911 and the fuel source encoder 2912) is connected to the input of the second data store.Attorney Docket No. FWIG-003W001

[0270] The 3D image generator 3001 is configured to generate a volumetrically rendered geospatially calibrated 3D reconstruction of the target building structure using the digital images that collectively represent all outside walls of the target building structure. The volumetrically rendered geospatially calibrated 3D reconstruction of the target building structure can be generated according to method 10 of Fig. 1.

[0271] The structural feature detector 2901, the fuel source detector 2902, the structural feature encoder 2911, and the fuel source encoder 2912 can each include a respective one or more processors 2930 that can run one or more respective trained ML models that can be stored in computer memory 2940 (e.g., volatile and / or non-volatile computer memory). The 3D image generator 2920 can also include one or more processors 2930 and computer memory 2940 that can store image data 2620 and / or computer-readable instructions.Alternatively, the structural feature detector 2901, the fuel source detector 2902, the structural feature encoder 2911, the fuel source encoder 2912, and / or the 3D image generator 2920 can share one or more processors 2930 and / or computer memory 2940.

[0272] Fig. 30 is a block diagram of the annotation engine 2704 of Fig. 26 according to one or more embodiments. The annotation engine 2704 can include a graphical display engine 3001, a graphical bounding engine 3002, a user input detector 3003, and an annotation engine 3004.

[0273] The graphical display engine 3001 is configured to produce graphics data that represents a portion of the volumetrically rendered geospatially calibrated 3D reconstruction of the target building structure that is output from the 3D image generator 2920 of Fig. 29. Additionally or alternatively, the graphical display engine 3001 is configured to cause a portion of the volumetrically rendered geospatially calibrated 3D reconstruction of the target building structure to be displayed on the display screen 2706 of Fig. 26. The portion of the volumetrically rendered geospatially calibrated 3D reconstruction of the target building structure to be displayed corresponds to and / or is defined by a position and FOV of a virtual inspector relative to the volumetrically rendered geospatially calibrated 3D reconstruction of the target building structure.

[0274] The graphical bounding engine 3002 is configured to produce bounding regions, such as the bounding region 2810 of Figs. 27 and 28, that are aligned with and / or define the perimeter of objects represented in the displayed portion of the volumetrically rendered geospatially calibrated 3D reconstruction of the target building structure that have beenAttorney Docket No. FWIG-003W001 tagged manually or automatically, such as the tagged object 2800 of Figs. 27 and 28. The bounding boxes are produced according to the respective indexed locations (e.g., the first annotation 2850A of Figs. 27 and 28) of the tagged objects stored in the second data store 2703 of Fig. 26. The graphical bounding engine 3002 can output image coordinates and / or bounding region data to the graphical display engine 3001 to display (e.g., overlay) one or more bounding regions on the displayed portion of the volumetrically rendered geospatially calibrated 3D reconstruction of the target building structure.

[0275] The user input detector 3003 is configured to detect user input relating to the graphics data displayed on the graphical display engine 3001, such as the displayed portion of the volumetrically rendered geospatially calibrated 3D reconstruction of the target building structure and / or the bounding regions. The user input detector 3003 can produce an output signal that represents a user input to move the position and / or FOV of the virtual inspector with respect to the displayed portion of the volumetrically rendered geospatially calibrated 3D reconstruction of the target building structure. This output signal can be sent to the graphical display engine 3001 and can cause the graphical display engine 3001 to update the displayed portion of the volumetrically rendered geospatially calibrated 3D reconstruction of the target building structure shown on the display screen. Updating the displayed portion of the volumetrically rendered geospatially calibrated 3D reconstruction of the target building structure can be performed according to method 70 of Fig. 6 in one or more embodiments.

[0276] The user input detector 3003 can also produce an output signal that represents a user input to adjust a position, dimension(s), and / or shape (e.g., one or more vertices) of a bounding region for an object. This output signal can be sent to the graphical bounding engine 3002 to cause a corresponding update to the position, dimension(s), and / or shape of the bounding region as graphically displayed on the display screen. The graphical bounding engine 3002 can generate a corresponding output signal to update the indexed locations of the tagged objects stored in the second data store. Alternatively, the graphical bounding engine 3002 can generate a corresponding output signal to the annotation engine 3004 to update the indexed locations of the tagged objects stored in the second data store, as further described below.

[0277] The user input detector 3003 can also produce an output signal that represents a user input to tag an object represented in the displayed portion of the volumetrically rendered geospatially calibrated 3D reconstruction of the target building structure. This output signal can be sent to the annotation engine 3004 which can cause the annotation engine to generateAttorney Docket No. FWIG-003W001 one or more annotation fields (e.g. GUI elements 1210 of Fig. 11 and / or annotations 2850 of Figs. 27 and 28) that can be displayed on the display screen by the graphical display engine 3001. The data for the annotation field(s) can be received or detected by the user input detector 3003 and entered by the annotation engine 3004. When the user input detector 3003 detects that a user input representing a completion of the annotation(s) for the tagged object, the annotation engine 3004 can send the annotation data to the second data store 2703 of Fig.26 to add or update the encoded fuel descriptors or the encoded structural descriptors stored in the second data store 2703 according to the annotation data. The annotation data includes an indexed location for the tagged object corresponding to the position, dimension(s), and shape of the bounding region. An update of the position, dimension(s), and / or shape of the bounding region causes a corresponding update to the indexed location of the tagged object in the annotation data.

[0278] The graphical display engine 3001, the graphical bounding engine 3002, the user input detector 3003, and the annotation engine 3004 can each include a respective one or more processors 2930 and respective computer memory 2940 (e.g., volatile and / or nonvolatile computer memory). Alternatively, the graphical display engine 3001, the graphical bounding engine 3002, the user input detector 3003, and the annotation engine 3004 can share one or more processors 2930 and / or computer memory 2940.

[0279] Fig. 31 is a block diagram of the wildfire ignition risk modelling engine 2705 of Fig.26 according to one or more embodiments. The wildfire ignition risk modelling engine 2705 can include one or more risk processor 3100 and computer memory 3110 (e.g., non-volatile and / or volatile memory). A wildfire loss model 3120 is stored as computer-readable instructions in the computer memory 3110 that are configured to be executed by the risk processor(s) 3100. The wildfire risk loss model 3120 is configured to model the risk of wildfire ignition risk to a building structure. An example of the wildfire risk loss model 3120 is a PIM. The wildfire risk model 3120 receives as inputs the encoded fuel descriptors or the encoded structural descriptors stored in the second data store 2703 of Fig. 26.

[0280] Fig. 32 is a block diagram of a system and method 3200 for performing virtual inspections of a target building structure to determine a risk of loss to the building structure due to a wildfire according to one or more embodiments. The system and method 3200 include an input data store 3210, an image processor 3220, a virtual inspection engine 3230, a property features data store 3240, and a post processor 3250.Attorney Docket No. FWIG-003W001

[0281] The input data store 3210 includes computer memory that stores wrap-around image data 3212 geospatial footprint data 3214. The wrap-around image data 3212 includes digital images (video images and / or photographic images) that represent all outside walls of the target building structure. The digital images can include some or all of the property adjacent to the outside walls of the target building structure, such as within about 30 feet of the outside walls of the target building structure. In one or more embodiments, sensor data from the device that the captured the digital images can be included with the wrap-around image data 3212. For example, the device can be a mobile computing device such as smartphone that includes a magnetometer, a gyroscope, an accelerometer, and / or geospatial positioning circuitry. Sensor data from the magnetometer, the gyroscope, the accelerometer, and / or the geospatial positioning circuitry that is time-synchronized with the wrap-around image data 3212 can be used for alignment as discussed below. The input data store 3210 can comprise non-volatile computer memory that can be located in a computer or a server or distributed across one or more computers and / or servers.

[0282] The geospatial footprint data 3214 represents a geospatially calibrated footprint or perimeter (e.g., a geospatially calibrated polygon 420 of Fig. 4) of the target building structure that can be extracted from an geospatially calibrated overhead image of the property including the target building structure or can be provided from a third party.

[0283] An output of the input data store 3210 is coupled to an input of the image processor 3220. An output of the image processor 3220 is coupled to an input of the virtual inspection engine 3230 and to an input of the property features data store 3240. An output of the virtual inspection engine 3230 is coupled to an input of the property features data store 3240. An output of the virtual inspection engine 3230 and an output of the property features data store 3240 are coupled to an input of the post processor 3250. The property features data store 3240 can comprise non-volatile computer memory that can be located in a computer or a server or distributed across one or more computers and / or servers.

[0284] In one or more embodiments, the output of the virtual inspection engine 3230 can be coupled to an input of an optional diff processor 3260. The output of the optional diff processor 3260 is coupled to the input of the property features data store 3240. When the optional diff processor 3260 is included, the output of the virtual inspection engine 3230 is indirectly coupled to the input of the property features data store 3240 (via the optional diff processor 3260). When the optional diff processor 3260 is not included, the output of the virtual inspection engine 3230 can be directly coupled to the input of the property featuresAttorney Docket No. FWIG-003W001 data store 3240 or indirectly coupled to the input of the property features data store 3240 (but not via the optional diff processor 3260). The optional diff processor 3260 can be configured to generate differential data (or “diffs”) that can be applied to the existing property features data 3245 representing only the changes to be applied to the existing property features data 3245. The optional diff processor 3260 can be the same as the diff generator 4334 described below.

[0285] Fig. 33 is a block diagram of the image processor 3220 of Fig. 32 according to one or more embodiments. The image processor 3220 includes a 3D image generator 3310, an optional alignment engine 3320, and an optional image recognition engine 3330. The image processor 3220 can include one or more processors and / or computer memory (e.g., volatile and / or non-volatile computer memory). An output of the 3D image generator 3310 is coupled to an input of the optional alignment engine 3320 and to an input of the optional image recognition engine 3330.

[0286] The 3D image generator 3310 is configured to generate a volumetrically rendered 3D reconstruction of the target building structure using the wrap-around image data 3212. The volumetrically rendered 3D reconstruction of the target building structure can be generated according to method 10 of Fig. 1. In one or more embodiments, the time-synchronized sensor data can be included in the wrap-around image data 3212 used to generate volumetrically rendered 3D reconstruction of the target building structure. The sensor data can be used to scale the volumetrically rendered 3D reconstruction of the target building structure according to the units of the time-synchronized sensor data, for example by triangulation using known positions (from the sensor data) of the camera (or mobile computing device that includes a camera) when the digital images for the wrap-around image data 3212 were captured.Additionally or alternatively, the sensor data can be used to determine the geospatial position of the points of the volumetrically rendered 3D reconstruction of the target building structure.

[0287] The 3D image generator 3310 is configured to generate a volumetrically rendered 3D reconstruction of the target building structure in step 3301 and stores the volumetrically rendered 3D reconstruction of the target building structure in step 3302. Image data representing the volumetrically rendered 3D reconstruction of the target building structure can be stored in non-volatile computer memory such as in the property features data store 3240 of Fig. 32. Alternatively, a link to the image data representing the volumetrically rendered 3D reconstruction of the target building structure can be stored in the property features data store 3240 and the image data representing the volumetrically rendered 3DAttorney Docket No. FWIG-003W001 reconstruction of the target building structure can be stored in another location such as another data store.

[0288] The optional alignment engine 3320 receives as inputs the volumetrically rendered 3D reconstruction of the target building structure and the geospatial footprint data 3214 (Fig. 32). The optional alignment engine 3320 is configured to perform optional steps 3303 and 3304. In optional step 3303, the optional alignment engine 3320 can automatically align the volumetrically rendered 3D reconstruction of the target building structure and the geospatial footprint data 3214. Optional step 3303 can be performed using the geospatial position of the points of the volumetrically rendered 3D reconstruction of the target building structure, which can be determined using the sensor data as discussed above. Additionally or alternatively, optional step 3303 can be performed using known geospatial locations of visible objects or visible markers that are represented in the wrap-around image data 3212. The output(s) of optional step 3303 is / are alignment data that can include or consist of a scale factor and a transformation matrix that can be applied to the volumetrically rendered 3D reconstruction of the target building structure (and any surrounding property captured in the wrap-around image data 3212 of Fig. 32) so as to align the volumetrically rendered geospatially calibrated 3D reconstruction of the target building structure with the geospatial footprint data 3214 to form a volumetrically rendered geospatially calibrated 3D reconstruction of the target building structure (and any surrounding property captured in the wrap-around image data 3212 of Fig. 32). These alignment data can be stored in optional step 3304 such as in the property features data store 3240 of Fig. 32.

[0289] The alignment data can be persisted such that a volumetrically rendered geospatially calibrated 3D reconstruction of the target building structure (and any surrounding property captured in the wrap-around image data 3212 of Fig. 32) can be generated and / or accessed in the future without having to perform an automatic alignment in optional step 3303.

[0290] The optional image recognition engine 3330 receives as an input the volumetrically rendered geospatially calibrated 3D reconstruction (or the volumetrically rendered 3D reconstruction) of the target building structure. The optional image recognition engine 3330 can be configured to detect structural features and / or insurance policy violations, such as debris, wood piles, and / or other fuel sources and / or the lack of certain represented (or not represented) in the volumetrically rendered geospatially calibrated 3D reconstruction (or the volumetrically rendered 3D reconstruction) of the target building structure and / or in the surrounding property represented therein. The optional image recognition engine 3330 canAttorney Docket No. FWIG-003W001 include one or more trained ML models that is / are configured to detect structural features and / or insurance policy violations.

[0291] The optional image recognition engine 3330 is configured to perform optional steps 3305 and 3306. In optional step 3305, objects, such as structural features and / or insurance policy violations, are automatically detected (e.g., using one or more trained ML models) in the volumetrically rendered geospatially calibrated 3D reconstruction (or the volumetrically rendered 3D reconstruction) of the target building structure and / or in the wrap-around image data 3212. Automatic detection in step 3305 can include detecting a type of each detected object (e.g., structural feature and / or insurance policy violation), one or more properties or materials of each detected object, and indexing the position and region / bounds of the detected object in image space.

[0292] In optional step 3306, the indexed position and region / bounds of each object detected in step 3305 is mapped to the image space of the volumetrically rendered geospatially calibrated 3D reconstruction (or the volumetrically rendered 3D reconstruction) of the target building structure. The indexed position and region / bounds for a given detected object can be represented as a series of coordinates in image space that define a two-dimensional shape such as a polygon that corresponds to the indexed position and region / bounds for that object. The two-dimensional shape such as a polygon can be the same as the bounding region 2810 of Figs. 27 and 28. In some embodiments, image recognition can be used to determine the location of objects. The coordinates of those objects would be in the three-dimensional image space. Using the transformation matrix created during the alignment and leveling step, the geospatial coordinates can be calculated with the x,z parts of the 3D coordinates. Height data can also be extracted from the y coordinate if the ground level is calibrated.

[0293] The image processor 3220 can also be configured to perform optional steps 3307 and 3308. Optional steps 3307 and 3308 can be performed when optional steps 3305 and 3306 are also performed. In optional step 3307, the image processor 3220 can generate bounding regions, such as the bounding region 2810 of Figs. 27 and 28, that can be overlaid and / or added to the image data representing the volumetrically rendered geospatially calibrated 3D reconstruction (or the volumetrically rendered 3D reconstruction) of the target building structure as stored in data store such as the property features data store 3240 of Fig. 32.

[0294] In optional step 3308, the image processor 3220 can send annotation data representing the detected objects to the property features data store 3240 of Fig. 32 that can store theAttorney Docket No. FWIG-003W001 annotation data as property features data 3245, as shown in Fig. 32. The annotation data representing the detected objects can include for each detected object: a unique identifier, the indexed position and region / bounds (e.g., as represented as coordinates in image space), the object type, and / or any detected properties / materials. When automatic alignment is performed in optional step 3303, the transformation matrix can be applied to the indexed position and region / bounds of each tagged object such that the indexed position and region / bounds of each tagged object are properly aligned with the aligned volumetrically rendered geospatially calibrated 3D reconstruction of the target building structure. In addition, the geospatial footprint data 3214 and the indexed position and region / bounds of each tagged object are properly aligned with the aligned volumetrically rendered geospatially calibrated 3D reconstruction of the target building structure can be used to determine a corresponding geospatial position of each tagged object.

[0295] The property features data store 3240 of Fig. 32 can include non-transitory computer memory that stores tagged objects and corresponding annotation data as property features data 3245 in a file, a database, a table, and / or another form or structure. In one or more embodiments, the property features data store 3240 of Fig. 32 can store the tagged objects as objects in a JavaScript Object Notation (JSON) file and the corresponding annotation data can be stored as respective properties or values of the objects in the JSON file. The property features data store 3240 can be the same as or different than the second data store 2703 of Fig. 26.

[0296] Fig. 34 is a block diagram of the virtual inspection engine 3230 of Fig. 32 according to one or more embodiments. The virtual inspection engine 3230 includes a rendering engine 3410, a tagged object rendering engine 3420, a user input detector 3430, an optional alignment engine 3440, an optional auto-detection review engine 3450, and a visual inspection engine 3460. The virtual inspection engine 3230 can include one or more processors and / or computer memory (e.g., volatile and / or non-volatile computer memory). The virtual inspection engine 3230 can be implemented in a mobile computing device such as a smartphone, tablet, or laptop, such as through a dedicated application, in one or more embodiments. Alternatively, the virtual inspection engine 3230 can be in communication with a virtual inspection application 3270 that can run on a mobile computing device such as a smartphone, tablet, or laptop, as shown in Fig. 32. The output of the virtual inspection application 3270 can be coupled directly or indirectly to the input of the diff processor 3260. The output of the virtual inspection application 3270 can be indirectly coupled to the input ofAttorney Docket No. FWIG-003W001 the diff processor 3260 via the virtual inspection engine 3230. In one or more embodiments, the virtual inspection engine 3230 can be the same as the annotation engine 2704 of Fig. 26.

[0297] The rendering engine 3410 is configured to receive, from the property features data store 3240 of Fig. 32, image data representing the volumetrically rendered geospatially calibrated 3D reconstruction (or the volumetrically rendered 3D reconstruction) of the target building structure. The rendering engine 3410 can also be configured to receive image data representing a geospatially calibrated aerial (or overhead) view of the target property including the target building structure. The aerial view can be represented in two dimensions.

[0298] In step 3400, the rendering engine 3410 produces output data that represents a portion of the volumetrically rendered geospatially calibrated 3D reconstruction (or the volumetrically rendered 3D reconstruction) of the target building structure for display on a display screen (e.g., on the mobile computing device or coupled to another computing device). The displayed portion of the volumetrically rendered geospatially calibrated 3D reconstruction (or the volumetrically rendered 3D reconstruction) of the target building structure corresponds to a position and FOV of a virtual inspector or of a virtual camera relative to the volumetrically rendered geospatially calibrated 3D reconstruction (or the volumetrically rendered 3D reconstruction) of the target building structure. The portion of the volumetrically rendered geospatially calibrated 3D reconstruction (or the volumetrically rendered 3D reconstruction) of the target building structure that is displayed initially can represent a default position and / or a default FOV of the virtual inspector or camera.

[0299] In optional step 3401, the rendering engine 3400 produces output data that represents a geospatially calibrated aerial (or overhead) rendering of the target property including the target building structure for display on the display screen. The aerial rendering can comprise a graphically rendering aerial image. Alternatively, the aerial rendering can be replaced or supplemented with an aerial photographic image such as a satellite image. The aerial rendering includes any fuel sources on the target property such as trees (or other large vegetation) and any secondary building structures (e.g., a shed, a detached garage, a guest house, a pool house, etc.). The aerial rendering of the target property can be displayed on the display screen separately or simultaneously with the portion of the volumetrically rendered geospatially calibrated 3D reconstruction (or the volumetrically rendered 3D reconstruction) of the target building structure. When displayed simultaneously, the aerial rendering can be overlaid with (or over) the 3D reconstruction. The aerial rendering of the target property can be derived or generated from the geospatially calibrated overhead image 2200 of Fig. 21.Attorney Docket No. FWIG-003W001

[0300] The tagged object rendering engine 3420 is configured to receive, from the property features data store 3240 of Fig. 32, the property features data 3245 including the indexed image locations of the tagged objects with respect to the volumetrically rendered geospatially calibrated 3D reconstruction (or the volumetrically rendered 3D reconstruction) of the target building structure and the indexed image locations of the tagged objects with respect to the aerial rendering. In step 3402, the tagged object rendering engine 3420 generates bounding regions (or bounding polygons), such as a bounding region 2810 of Figs. 27 and 28, at the indexed image locations of the tagged objects and overlaid on the volumetrically rendered geospatially calibrated 3D reconstruction (or the volumetrically rendered 3D reconstruction) of the target building structure and / or on the aerial rendering. The tagged objects include any objects that were automatically detected by the optional image recognition engine 3330 and any objects that are manually tagged with the visual inspection engine 3460, as discussed below.

[0301] The tagged object rendering engine 3420 can display the bounding region(s) that is / are viewable in the displayed portion of the volumetrically rendered geospatially calibrated 3D reconstruction (or the volumetrically rendered 3D reconstruction) of the target building structure, which can be provided from the rendering engine 3410 as an input to the tagged object rendering engine 3420. Additionally or alternatively, the tagged object rendering engine 3420 can display the bounding region(s) at the corresponding image locations in the aerial rendering even though a tagged object may not be viewable from the perspective of the aerial rendering. For example, the tagged object rendering engine 3420 can display the bounding regions for windows on the outside walls of the building structure in the aerial rendering even though these windows are not viewable in the aerial view.

[0302] In one or more embodiments, the rendering engine 3410 and the tagged object rendering engine 3420 can be combined as a rendering engine 3415. The rendering engine 3415 can include both the rendering engine 3410 and the tagged object rendering engine 3420 (in which case the rendering engine 3410 and the tagged object rendering engine 3420 can be referred to a 3D rendering sub-engine 3410 and a tagged object rendering sub-engine 3420, respectively) and can be configured to perform steps 3400-3402. Alternatively, the rendering engine 3415 can be configured to perform steps perform steps 3400 and 3402 but may not have include distinct rendering engines or rendering sub-engines for 3D rendering and / or for tagged object rendering.Attorney Docket No. FWIG-003W001

[0303] Fig. 35 A shows a geospatially calibrated aerial (or overhead) rendering 3500 of a target property 3510 including a target building structure 3520. The aerial rendering 3500 shows trees (or other large / major vegetation) 3530 and a secondary building structure 3540. The secondary building structure 3540 can be or comprise a shed, a detached garage, a guest house, a pool house, or another ancillary building structure. The aerial image 3500 can also include a property boundary line 3550. In one or more embodiments, the aerial image 3500 can further include neighboring properties 3560 and neighboring building structures 3562, respectively.

[0304] Bounding regions 3570 for tagged objects 3572 can be displayed on the aerial rendering 3500. For example, bounding regions 3570 can be displayed on the trees 3530 and on the perimeter of the building structure 3540 at the indexed locations for any structural features 3580 that have been tagged (as tagged objects 3572) manually and / or automatically and stored in the property features data store 3240 of Fig. 32. For example, bounding regions 3570 for windows on the perimeter of the building structure 3540 can be displayed at image locations corresponding to indexed geospatial locations for those structural features 3580.

[0305] Fig. 35B shows an overlay 3580 of the geospatially calibrated aerial rendering 3500 of Fig. 35 A and a portion of a volume-rendered reconstruction 3590 of the target building structure 3520 according to one or more embodiments. The volume-rendered reconstruction 3590 can be the same as the volume-rendered reconstruction 60 of Fig. 5. The overlay 3580 can allow a virtual inspector and / or a service provider to view where other objects, such as trees 3530 and respective bounding regions 3570, are relative to the displayed portion of the geospatially calibrated volume-rendered reconstruction 3590.

[0306] Returning to Fig. 34, the user input detector 3430 is configured to detect, in step 3403, user input corresponding to a change in position of and / or change in FOV of a virtual inspector (e.g., a virtual inspector 800 of Figs. 7A, 8 A, and 9A) relative to the displayed portion of the volumetrically rendered geospatially calibrated 3D reconstruction (or the volumetrically rendered 3D reconstruction) of the target building structure or a change in position and / or change in FOV of a virtual camera relative to the displayed portion of the volumetrically rendered geospatially calibrated 3D reconstruction (or the volumetrically rendered 3D reconstruction) of the target building structure. The user input detector 3430 can produce an output representing or corresponding to a detected change in position of and / or a detected change in FOV of a virtual inspector or camera relative to the displayed portion of the volumetrically rendered geospatially calibrated 3D reconstruction (or the volumetricallyAttorney Docket No. FWIG-003W001 rendered 3D reconstruction) of the target building structure. This output from the user input detector 3430 is provided as an input to the rendering engine 3410 which can trigger the rendering engine 3410 to repeat step 3400 so as to update the rendering of the displayed portion of the volumetrically rendered geospatially calibrated 3D reconstruction (or the volumetrically rendered 3D reconstruction) of the target building structure according to the user input. The update to the rendering of the displayed portion of the volumetrically rendered geospatially calibrated 3D reconstruction (or the volumetrically rendered 3D reconstruction) of the target building structure and / or the output from the user input detector 3430 is / are provided as an input to the tagged object rendering engine 3420. This input to the tagged object rendering engine 3420 can trigger to the tagged object rendering engine 3420 to repeat step 3402 so as to update the bounding region(s) that is / are viewable in the updated rendering of the displayed portion of the volumetrically rendered geospatially calibrated 3D reconstruction (or the volumetrically rendered 3D reconstruction) of the target building structure. Steps 3400-3403 can be performed iteratively in a loop, for example according to step 1401 of Fig. 13, to allow a user to virtually walk or move about the volumetrically rendered geospatially calibrated 3D reconstruction (or the volumetrically rendered 3D reconstruction) of the target building structure to virtually inspect a property and building structure for purposes of evaluating wildfire ignition risks to the building structure and / or inspecting for wildfire insurance policy violations.

[0307] The optional alignment engine 3440 is configured to provide a user interface for performing manual alignment of the volumetrically rendered 3D reconstruction of the target building structure. The optional alignment engine 3440 is configured to perform optional steps 3404-3406, which can be performed before or after any of steps 3400-3403 or before or after any of steps 3407-3411, discussed below.

[0308] In optional step 3404 and referring to Figs. 36A, an outlined two-dimensional projection 3600 of the volumetrically rendered 3D reconstruction 3602 of the target building structure and an outlined two-dimensional projection 3610 of the geospatial footprint or ground plane 3612 are simultaneously displayed. The outlined two-dimensional projections 3600 and 3610 are not aligned in Fig. 36A.

[0309] Returning to Fig. 34 and referring to Figs. 36A and 36B, in optional step 3405, the building structure project! on / outline 3600 of the volumetrically rendered 3D reconstruction 3602 is graphically repositioned, in response to one or more user inputs, with respect to the outlined two-dimensional projection 3610 of the geospatial footprint or ground plane 3612 soAttorney Docket No. FWIG-003W001 as to visually align the outlined two-dimensional projections 3600 and 3610, as shown in Fig.36B. Optional step 3405 may be repeated iteratively until visual alignment is completed.

[0310] When visual alignment is completed, as indicated by a user input, in optional step 3406, alignment data for the volumetrically rendered 3D reconstruction of the target building structure is stored as or with the property features data 3245 in the property features data store 3240 of Fig. 34. The alignment data can include or consist of a scale factor and a transformation matrix. When applied to the sparse point cloud for the volumetrically rendered 3D reconstruction of the target building structure, the alignment data transforms the volumetrically rendered 3D reconstruction of the target building structure (and any surrounding property captured in the wrap-around image data 3212 of Fig. 32) to a volumetrically rendered geospatially calibrated 3D reconstruction of the target building structure (and any surrounding property captured in the wrap-around image data 3212 of Fig.32). The alignment data stored in the property features data store 3240 can be persisted such that a volumetrically rendered geospatially calibrated 3D reconstruction of the target building structure (and any surrounding property captured in the wrap-around image data 3212 of Fig.32) can be generated and / or accessed in the future without having to perform a visual alignment, in optional steps 3404 and 3505, of the outlined two-dimensional projections 3600 and 3610.

[0311] Continuing with Fig. 34, the auto-detection review engine 3450 is configured to allow a user to manually review any tagged objects including any objects that were automatically detected by the optional image recognition engine 3330 of Fig. 33. The auto-detection review engine 3450 is configured to perform optional steps 3407-3410.

[0312] In optional step 3407, the auto-detection review engine 3450 causes any tagged objects that were automatically detected by the optional image recognition engine 3330 to be displayed in the displayed portion of the volumetrically rendered 3D reconstruction of the target building structure (which may or may not be geospatially calibrated) and / or in the geospatially calibrated aerial image. In one or more embodiments, only these automatically detected tagged objects are displayed (e.g., any manually tagged objects are not displayed simultaneously with the automatically detected tagged objects). Alternatively, the visual attributes of the automatically detected tagged objects can visually different than the visual attributes of any manually tagged objects. Displaying the automatically detected tagged objects includes displaying their respective two-dimensional bounding regions in an overlay over the volumetrically rendered 3D reconstruction of the target building structure.Attorney Docket No. FWIG-003W001

[0313] In optional step 3408, one or more of the bounding regions for the automatically detected tagged objects can be manually edited in response to one or more user inputs. A user may wish to edit a bounding region for an automatically detected tagged object when the bounding region does not conform (e.g., in shape or position) to the shape of the automatically detected tagged object. Updating the bounding region for an automatically detected tagged object causes a corresponding update to the annotation data for the indexed image location of the automatically detected tagged object and, when the volumetrically rendered 3D reconstruction is a volumetrically rendered geospatially calibrated 3D reconstruction, an indexed geospatial location of the automatically detected tagged object. In addition to providing a geospatial location, the indexed geospatial location can provide dimensional measurements of the automatically detected tagged object.

[0314] In optional step 3409, one or more annotations can be added or edited, in response to one or more user inputs, for at least one (or some or all) of the automatically detected tagged objects. The annotation(s) added or edited can represent a type or category of an automatically detected tagged object, a material(s) of the automatically detected tagged object, a location of an automatically detected tagged object, a mitigation action(s) for the automatically detected tagged object, a note(s) for the automatically detected tagged object, and / or another annotation(s).

[0315] In optimal step 3410, one or more automatically detected tagged objects can be removed in response to one or more user inputs. A detected tagged object can be removed, for example, when the user determines that the detected tagged object was improperly tagged.

[0316] Steps 3407-3410 can be repeated iteratively as the virtual inspector or camera “moves” about the volumetrically rendered 3D reconstruction of the target building structure (and any surrounding property) according to user inputs (e.g., according to optional steps 3400-3403).

[0317] The visual inspection engine 3460 is configured to allow a user to manually tag objects represented in the volumetrically rendered 3D reconstruction of the target building structure (which may or may not be geospatially calibrated) and / or in the geospatially calibrated aerial image of the target property. The visual inspection engine 3460 is configured to perform steps 3411-3413.

[0318] In step 3411, one or more objects represented in the volumetrically rendered 3D reconstruction of the target building structure and / or in the geospatially calibrated aerialAttorney Docket No. FWIG-003W001 image of the target property is / are manually tagged. An object can be manually tagged in response to one or more user inputs. With respect to the volumetrically rendered 3D reconstruction of the target building structure (and any surrounding property), object(s) can be manually tagged as the virtual inspector or camera “moves” about the volumetrically rendered 3D reconstruction of the target building structure (and any surrounding property) according to user inputs (e.g., according to optional steps 3400-3403).

[0319] In step 3412, a two-dimensional bounding region for each manually tagged object is defined. The two-dimensional bounding region can be graphically represented in an overlay over the displayed portion of the volumetrically rendered 3D reconstruction of the target building structure and / or over the geospatially calibrated aerial image of the target property. Defining a bounding region for a manually detected tagged object defines an indexed image location of the manually tagged object and, when the volumetrically rendered 3D reconstruction is a volumetrically rendered geospatially calibrated 3D reconstruction, an indexed geospatial location of the manually tagged object. Defining a bounding region for a manually detected tagged object in a geospatially calibrated aerial image of the target property can also define an indexed image location and an indexed geospatial location of the manually tagged object. In addition to providing a geospatial location, the indexed geospatial location can provide dimensional measurements of the manually tagged object. In one or more embodiments, defining a two-dimensional bounding region for an object in step 3412 automatically tags the object in step 3411.

[0320] In step 3413, annotation data for each tagged object is received in response to one or more user inputs. The annotation data can represent a type or category of an automatically detected tagged object, a material(s) of the automatically detected tagged object, a location of an automatically detected tagged object, a mitigation action(s) for the automatically detected tagged object, a note(s) for the automatically detected tagged object, and / or another annotation(s).

[0321] After steps 3301-3313 are completed, including or excluding any optional steps, in step 3314 the virtual inspection engine 3230 updates the property features data 3245 in the property features data store 3240 of Fig. 32 with any changes to tagged objects (whether manually tagged or automatically tagged) including any changes the respective annotation data compared to the current property features data 3245. In one example, these changes can be provided as a differential file (or “diffs”) that can be applied to an existing JSON file to form a new or updated version of the JSON file that includes the changes.Attorney Docket No. FWIG-003W001

[0322] Fig. 37 is a block diagram of the post processor 3250 of Fig. 32 according to one or more embodiments. The post processor 3250 can include a wildfire ignition risk modelling engine 3710 and a mitigation engine 3720.

[0323] The wildfire ignition risk modelling engine 3710 is configured to perform optional step 3701 to model a risk of loss to the building structure due to a wildfire, for example according to a PIM or another model. The property features data 3245 in the property features data store 3240 of Fig. 32 are provided as inputs to the wildfire ignition risk modelling engine 3710. The wildfire ignition risk modelling engine 3710 can be the same as or different than the wildfire ignition risk modelling engine 2705 of Figs. 26 and 31.

[0324] The mitigation engine 3720 is configured to perform optional step 3702 to generate service work orders that can be sent to a wildfire mitigation service provider and / or to the owner of the building structure. The property features data 3245 in the property features data store 3240 of Fig. 32 are provided as inputs to the mitigation engine 3720. The service work orders can correspond to any tagged objects that include mitigation actions and / or wildfire insurance policy violations in the annotation data.

[0325] Fig. 38 is a flow chart of a method 3800 for generating a geospatially calibrated aerial (or overhead) rendering of a target property according to one or more embodiments. Method 3800 can be used to generate the geospatially calibrated aerial image 3500 of Figs. 35 A and 35B according to one or more embodiments. Method 3800 can be performed by the image processor 3220 of Fig. 32 in one or more embodiments.

[0326] In step 3801 of Fig. 38, a geospatially calibrated aerial image that includes the target property is received. The geospatially calibrated aerial image can be or comprise a satellite image or another overhead digital photographic image. The geospatially calibrated aerial image has a known scale and includes geospatial data corresponding to the property(ies) represented in the image.

[0327] In step 3802, the target property is identified in the geospatially calibrated aerial image received in step 3801. The target property can be identified based on the geospatial data corresponding to the address of the target property, which can be determined using one or more third-party services.

[0328] In step 3803, a target building structure on the target property is identified in the geospatially calibrated aerial image. The target building structure can be the primary building structure on the target property, such as a residential house on a property that can include oneAttorney Docket No. FWIG-003W001 or more ancillary building structures such as a shed, a detached garage, a guest house, a pool house, or another ancillary building structure. The target building structure can be identified as the largest building structure on the target property.

[0329] In step 3804, a footprint for the target building structure is determined. The footprint can be determined using a trained ML model that is configured to detect building structures and / or the perimeters of building structures.

[0330] In step 3805, coordinates that represent vertices defining a geospatially calibrated polygon corresponding to the footprint for the target building structure are generated. Fig. 4 is an example of a geospatially calibrated polygon 420 where the coordinates represent vertices 430.

[0331] Returning to Fig. 38, in step 3806, trees and / or other large / major vegetation are detected in the geospatially calibrated aerial image. The trees and / or other large / major vegetation can be detected using a trained ML model that is configured to detect trees and / or large / major vegetation. Alternatively, a trained ML model that is configured to detect both (a) building structures and / or the perimeters of building structures and (b) trees and / or large / major vegetation can be used in steps 3804 and 3806.

[0332] In step 3807, data that defines geospatially calibrated bounding regions for the trees and / or other large / major vegetation detected in step 3806 are generated. The geospatially calibrated bounding regions can be formed as circles or other shapes. When the geospatially calibrated bounding regions are defined as circles, the data defining the geospatially calibrated bounding regions can include a center and a radius (or diameter) for each circle. When the geospatially calibrated bounding regions are defined as another shape, the data defining the geospatially calibrated bounding regions can include coordinates that represent each geospatially calibrated bounding region.

[0333] In step 3808, indexed locations for the geospatially calibrated polygon and the geospatially calibrated bounding regions are stored in a property features data store such as the property features data store 3240 of Fig. 32.

[0334] Fig. 39 is a block diagram of a virtual wildfire inspection apparatus 3900 according to one or more embodiments. The virtual wildfire inspection apparatus 3900 can also be referred to as a wildfire service provider apparatus 3900. The apparatus 3900 includes one or more microprocessors 3910, computer memory 3920, a display screen 3930, and user input device(s) 3940. The computer memory 3920 includes non-volatile memory that storesAttorney Docket No. FWIG-003W001 computer-readable instructions for performing the functions of the virtual inspection engine 3230 of Figs. 32 and 34 including some or all of steps 3400-3414 of Fig. 34. The microprocessors 3910 can perform one or more functions and / or one or more steps 3400-3414 of Fig. 34 automatically and / or in response to user input that can be received from the user input device(s) 3940. The user input device(s) 3940 can include a mouse, a touch pad, a keyboard, a touch screen, a virtual keyboard, a microphone, and / or a digital camera.

[0335] In one or more embodiments, the virtual inspection apparatus 3900 can include optional geospatial circuitry 3950 that can determine the geospatial coordinates of a service provider (or other user) while performing work (e.g., mitigation actions and / or insurance policy violation remedy actions) on site, such as for tracking and / or corroboration that the work was completed. In one or more embodiments, the virtual inspection apparatus 3900 (e.g., via the display screen) can prompt the service provider (or other user) to capture images of mitigation actions, insurance policy violations, insurance policy violation remedy actions, and / or other actions / objects using a digital camera in the user input device(s) 3940.

[0336] The virtual inspection apparatus 3900 can comprise a mobile computing device such as a smartphone, a tablet, or a laptop in one or more embodiments.

[0337] In one or more embodiments, the virtual inspection apparatus 3900 can include communications circuitry 3960 that can allow the virtual inspection apparatus 3900 to be in communication (e.g., network communication and / or local communication) with other devices and / or servers and / or with a network. For example, the virtual inspection apparatus 3900 can be in communication with the property features data store 3240, the image processor 3220, and / or the post processor 3250 of Fig. 32.

[0338] The virtual inspection apparatus 3900 can store at least some of the property features data 3245 locally in the computer memory 3920 such that the virtual inspection apparatus 3900 can operate independently such as in remote location where a wireless network (e.g., a cellular network, a WiFi network, and / or another network) may not be available. The property features data 3245 stored locally in the computer memory 3920 can include (a) image data representing the volumetrically rendered geospatially calibrated 3D reconstruction (or the volumetrically rendered 3D reconstruction) of the outside walls of the target building structure including structural features on the outside walls, (b) optional image data representing a geospatially calibrated aerial (or overhead) rendering of the target property including the target building structure, (c) tagged object data representing one or more taggedAttorney Docket No. FWIG-003W001 objects (including their respective indexed locations (geospatial and image locations)) and respective annotation data. The image data representing the volumetrically rendered geospatially calibrated 3D reconstruction (or the volumetrically rendered 3D reconstruction) of the outside walls of the target building structure can include at least a portion of the target property such as the target property adjacent to the outside walls of the target building structure. When the computer memory 3920 stores a volumetrically rendered 3D reconstruction that has not been geospatially aligned and / or registered, the computer memory 3920 can also store a geospatially calibrated footprint or polygon for the target building structure to use for such geospatially alignment and / or registration (e.g., which can be performed according to steps 104 and 105 of Fig. 1).

[0339] During virtual inspection or work by a service provider, updates to the stored property features data 3245 can be generated (e.g., in response to user inputs) and stored locally in the computer memory 3920, such as through a differential file or “diffs,” which can be sent to the property features data store 3240 of Fig. 32 using the communications circuitry 3960 when communication therebetween is or becomes available.Example Property Features Data Structure

[0340] Fig. 40 is a block diagram of a property features data structure 4000 according to some embodiments. In some embodiments, the property features data structure 4000 may be stored as a JSON file, which may use WGS84 / EPSG:4326 coordinate reference system for geospatial data. In some embodiments, the property features data structure 4000 aggregates property-related data for wildfire risk assessment and fire protection planning. In some embodiments, the property features data structure 4000 may employ 2.5D modeling, representing primarily geospatial latitude / longitude coordinates with associated height attributes, while three-dimensional renderings may serve as visualization aids for virtual inspection. In some embodiments, the property features data structure 4000 may be validated using a Pydantic model serialization framework.

[0341] In some embodiments, the property features data structure 4000 includes an info section 4010 that stores property information metadata. In some embodiments, the info section 4010 may include sub-fields 4012 such as an address (property street address), a center of analysis (representing latitude / longitude coordinates for the primary analysis point), risk data aggregated from multiple authoritative sources, client overrides for custom client configurations, and a protection plan. In some embodiments, the risk data may include valuesAttorney Docket No. FWIG-003W001 from CPUC (California Public Utilities Commission), CalFire (California Department of Forestry and Fire Protection), USDA (US Department of Agriculture), and other sources. In some embodiments, each risk value includes metadata such as a numeric risk value, risk label, source identifier, load version (date / time of data source), responsibility area type (such as SRA or Local Responsibility Area, LRA), selected hazard (the specific hazard factor such as dense tree canopy or proximity to fuel sources), and surveyed hazards. In some embodiments, when a property is located in an area where multiple data sources provide fire risk assessments, the system stores all risk values rather than selecting a single value, enabling risk aggregation and comparison across sources. In some embodiments, risk data is automatically refreshed on a scheduled basis (for example, quarterly or when authoritative sources update their assessments), and the system notifies property owners if risk levels change significantly (such as by more than one risk category). In some embodiments, the system maintains a version history of all risk data, enabling auditors to verify which risk values were in effect at the time a property assessment was conducted.

[0342] In some embodiments, the property features data structure 4000 includes structures data 4020 for building structures on the property. In some embodiments, each structure may be classified as either a main structure (representing the primary building structure to be protected) or one or more fuel structures (representing neighboring structures or outbuildings that may contribute to wildfire ignition risk). In some embodiments, only one main structure may be present for a given property.

[0343] In some embodiments, the property features data structure 4000 includes trees data 4030 for vegetation on or near the property. In some embodiments, tree data may be initially populated by a machine learning model trained on satellite imagery, with the model detecting tree positions, estimating canopy radius, height, and species. In some embodiments, tree data may be subsequently corrected by human operators using a property application. In some embodiments, the trees data 4030 is used by a property ignition model (PIM) to calculate wildfire ignition risk and defensible space requirements.

[0344] In some embodiments, the property features data structure 4000 includes an images section 4040 that stores image data. In some embodiments, the images section 4040 may include sub-fields 4042 for satellite imagery (including standard RGB, near-infrared NIR, and digital terrain / surf ace model DTM / DSM variants) and ground-level imagery. In some embodiments, satellite images may include metadata such as GPS coordinates, capture timestamp, resolution, provider information, scale factor, and shape bounds. In someAttorney Docket No. FWIG-003W001 embodiments, a three-dimensional model generated from wrap-around video using structure from motion and Gaussian splatting techniques may be linked as an image in the images section 4040.

[0345] In some embodiments, the property features data structure 4000 includes an elevation section 4050 that stores terrain elevation data. In some embodiments, the elevation section 4050 may include sub-fields 4052 defining geographic bounds (e.g., latitude and longitude bounds), grid dimensions (e.g., row and column counts), and elevation values (cells) representing elevation above sea level in meters. In some embodiments, the elevation data may be used by PIM to model slope-based fire spread behavior, as hot air rises and affects how far embers travel and how flames propagate. In some embodiments, the elevation grid represents terrain elevation at regular intervals (for example, 10-meter spacing) across the property and surrounding area. In some embodiments, the system computes slope and aspect from the elevation grid, as wind-driven fires accelerate based on slope direction. In some embodiments, the elevation grid is provided to the property ignition model, which uses the slope and aspect data to multiply the fire spread rate computed for flat terrain by slopedependent and aspect-dependent acceleration factors based on fire behavior research. In some embodiments, trees positioned at significantly higher elevation than the primary structure are weighted as lower ignition risk, as heat and embers from such trees have reduced probability of reaching the structure.

[0346] In some embodiments, the property features data structure 4000 includes parcels data 4060 that stores land parcel boundary data including assessor’s parcel numbers (APN), parcel shapes, default wall materials, default roof materials, number of stories, foundation type, livable square meters, and year built.

[0347] In some embodiments, the property features data structure 4000 includes “other shapes” data 4070 that stores miscellaneous features not formally categorized in other sections, such as decks, fences, fuel sources, fire hydrants, site access routes, spray plans, and retardant container locations.

[0348] In some embodiments, the property features data structure 4000 includes observations data 4080 that stores annotated findings from virtual inspection.

[0349] In some embodiments, PIM input 4090 may be derived from the structures data 4020, the trees data 4030, and the elevation section 4050 to determine probability of destruction of the property due to wildfire. In some embodiments, the property features data structure 4000Attorney Docket No. FWIG-003W001 may also include optional fields such as risk value, PIM parameters (containing runtime parameters such as windspeed in m / s and ambient temperature in Kelvin), and change attribution tracking.

[0350] In some embodiments, the property features data structure serves as a primary input to the PIM, which may calculate the probability that a given structure will be destroyed by wildfire. In some embodiments, the accuracy of PIM’ s probability estimate may be sensitive to the accuracy of property features data, particularly for trees within the defensible space zone. In some embodiments, the virtual inspection system may provide data quality metrics for tree and structure measurements, indicating the estimated uncertainty based on the source of the measurement (such as automatically detected versus manually measured data). In some embodiments, uncertainty estimates may be provided to the PIM, which may use either a conservative risk estimate or compute a probability distribution of outcomes.Example Structure Data Model

[0351] Fig. 41 is a block diagram of a structure data model according to some embodiments. In some embodiments, a Structure object 4100 represents a building structure and includes fields for an internal identifier, a structure type 4102 (which may be the primary protected structure or neighboring structures), a shape defined as a polygon in geospatial coordinates representing the aerial-view outline of the structure, number of stories, wall materials, foundation type, eaves characteristics, gutters presence / type, a consider burnable flag for fire risk calculations, livable square meters, and a PIM identifier for cross-referencing to PIM entities.

[0352] In some embodiments, the Structure object 4100 includes walls data 4110 containing one or more wall objects 4120. In some embodiments, each wall object 4120 represents a wall segment of the building structure and includes an internal identifier, a linestring defining the wall in geospatial coordinates (e.g., representing a line segment of the structure’s perimeter), and spans data.

[0353] In some embodiments, each wall includes one or more wall-span objects 4130 that represent features or material variations along the wall. In some embodiments, each wall-span object 4130 includes a span type 4132 (e.g., window, door, vent, material, roof height, etc.), percentage start (which may be constrained to values 0-100), percentage end (which may be constrained to values 0-100), material designation, glass type (for windows, such as tempered or non-tempered), roof height at the wall segment, associated image references, andAttorney Docket No. FWIG-003W001 associated observation references. In some embodiments, the internal identifier is unique across all spans for a single wall.

[0354] In some embodiments, the Structure object 4100 also references a roof object 4150 that includes an internal identifier, a shape defined as a polygon, materials data for roofing material classifications, and features data. In some embodiments, the features data includes roof feature objects 4160 that represent roof features such as chimneys, skylights, and vents. In some embodiments, roof feature types 4162 are specified via a field and may include chimney, skylight, vent, and / or any other suitable feature type. In some embodiments, vent features may include a vent type such as ridge or offridge, and a vent subtype such as dormer, low profile, turbine, or other.

[0355] A wall span percentage representation 4140 illustrates how spans may be positioned along a wall linestring 4142. In some embodiments, representing wall spans as percentages of the wall linestring prevents orphan spans that would occur if spans were stored as independent linestrings, because the span positions automatically adjust when the wall shape is modified. In some embodiments, when a wall’s perimeter is modified due to corrected structural measurements or updated GIS data, the percentage positions automatically scale to maintain the correct relative positions of all spans on that wall. In some embodiments, this automatic adjustment prevents orphaned spans that would require manual relocation in systems storing spans as absolute coordinates. In some embodiments, the percentage representation enables continuous positioning of spans at any fractional percentage between 0 and 100, supporting sub-pixel precision in the three-dimensional rendering while consuming minimal storage compared to storing full coordinate tuples for each span endpoint. In some embodiments, an interpolation calculation is performed with double-precision floating-point arithmetic to ensure sub-meter accuracy in geospatial positioning even for very long walls in large properties. For example, a window span 4144 may have percent start of 13 and percent end of 23, a door span 4146 may have percent start of 40 and percent end of 55, and a vent span 4148 may have percent start of 78 and percent end of 85.Example Tree Data Model

[0356] Fig. 42 is a block diagram of a tree data model according to some embodiments. In some embodiments, a tree object 4200 represents a tree on or near the property and includes fields for an internal identifier, type, shape (e.g., a polygon representing the tree canopy footprint), height (e.g., from base to top of crown in meters), natural crown radius (e.g.,Attorney Docket No. FWIG-003W001 untrimmed canopy radius in meters), natural lowest branch height (e.g., untrimmed height of lowest branch in meters), breast height diameter, species, trim info, and additional data for additional schema-less attributes. In some embodiments, trim info may enable PIM to calculate defensible space and mitigation options.

[0357] In some embodiments, the breast height diameter 4226 represents the trunk diameter measured at a person’s breast / chest height, which is a standard forestry measurement (also known as DBH, diameter at breast height). In some embodiments, this measurement may be used to calculate heat production during combustion based on the inventors’ appreciation of correlation between trunk diameter and thermal output.

[0358] In some embodiments, tree species may be selected from a controlled vocabulary based on tree shape characteristics such as spherical, conical, or Y-shape, as shape characteristics affect fire behavior modeling.

[0359] In some embodiments, the system validates tree data to ensure measurements are physically plausible. For example, the system may enforce the constraint that height > natural lowest branch height > 0, as a tree’s lowest branch cannot be higher than the tree’s total height. In some embodiments, the natural crown radius may be constrained based on the tree’s height and species characteristics. In some embodiments, the breast height diameter correlates with the tree’s biomass and heat output during combustion, as the inventors have appreciated relationships between trunk diameter and thermal energy release. In some embodiments, the trim information may be used in conjunction with species characteristics to estimate how much a tree’s thermal contribution to fire behavior can be reduced through trimming.

[0360] In some embodiments, the tree object 4200 includes a tree time info object 4210 that stores trimming-related data for defensible space calculations. In some embodiments, the tree trim info object 4210 includes a max radius reduction (representing the maximum amount the canopy radius could be reduced by trimming), an actual trim height (representing the current trim height, which may be based on observed cut scars identified by an inspector), a max trim height (representing the maximum height to which the tree could possibly be trimmed), and a highest adjacent roofline (representing the height of the highest roofline of adjacent structures).

[0361] A tree measurement illustration 4220 shows the spatial relationships between tree measurements according to some embodiments. In some embodiments, the natural crownAttorney Docket No. FWIG-003W001 radius 4222 represents the untrimmed canopy extent. In some embodiments, the height 4224 represents the total tree height from ground level to the top of the canopy. In some embodiments, the breast height diameter represents the trunk diameter at breast height. In some embodiments, natural lowest branch height 4228 represents the height of the lowest branch above ground in the natural untrimmed state. In some embodiments, the highest adjacent roofline may be the roofline height of a nearby structure relative to the tree.

[0362] A defensible space calculation illustration 4240 shows how the trim information enables calculation of defensible space requirements according to some embodiments. In some embodiments, the max radius reduction 4242 indicates with dashed line circle the extent to which the tree canopy can be reduced (e.g., from the solid line circle) to achieve adequate clearance from the structure. In some embodiments, the trim information may be used to determine whether a tree needs to be removed entirely or can simply be trimmed to reduce wildfire ignition risk. In some embodiments, trees positioned above a structure (based on elevation data) may be assessed as lower risk than trees at the same elevation as the structure. The following figure will show a data flow that may be used with a property features data structure, followed by figures showing other details relating to a property features data structure.Example Property Features Data Flow

[0363] Fig. 43 is a flow diagram of a property features data flow according to some embodiments. In some embodiments, input sources 4300 include an address 4302, satellite imagery 4304, parcel data 4306, and structure outlines 4308 obtained from GIS data providers.

[0364] In some embodiments, a landscape pipeline 4310 receives the input sources and automatically generates an initial property features file. In some embodiments, the landscape pipeline 4310 may include a tree machine learning (ML) model 4312 that analyzes satellite imagery to detect trees and estimate their position, canopy radius, height, and species. In some embodiments, the landscape pipeline 4310 may also include geocoding 4314 to convert the address to latitude / longitude coordinates representing a center of analysis.

[0365] In some embodiments, an initial property feature file 4320 is generated by the landscape pipeline 4310 and stored in blob storage 4322 (e.g., as a JSON file). In some embodiments, the property feature file may be linked to a portfolio property database record for status tracking.Attorney Docket No. FWIG-003W001

[0366] In some embodiments, a property app 4330 (which may be referred to as a data repair tool, or a virtual inspection tool) provides a user interface for manual correction of the property features data. In some embodiments, the property app 4330 may receive a three-dimensional model 4332 generated from wrap-around video of the primary structure, captured using Structure-from-Motion (SfM) techniques. In some embodiments, the SfM pipeline employs feature-based visual odometry to establish consistent camera pose estimation across the image sequence, enabling accurate three-dimensional reconstruction even when imaging conditions vary due to changes in lighting, reflectance properties of building materials (such as reflective siding or windows), and weather conditions. In some embodiments, the use of multi -view stereo reconstruction in the context of building inspection for wildfire risk assessment represents a novel application, as the method automatically identifies and preserves spatial relationships between structural features (windows, doors, vents) and the building envelope, enabling precise measurement of feature dimensions and material types critical to fire ignition modeling. In some embodiments, the three-dimensional model 4332 enables virtual inspection and annotation of features not visible from aerial imagery, and may be stored in a Gaussian splatting format (such as .splat).

[0367] In some embodiments, the system captures time-synchronized inertial measurement unit (IMU) data from the mobile device during image acquisition, including magnetometer readings for orientation, gyroscope data for angular velocity, and accelerometer data for linear acceleration, such as is described herein. In some embodiments, these sensor signals enable the system to estimate camera pose and trajectory without requiring manually identified ground control points or known reference features. In some embodiments, the trajectory information is used in conjunction with the sparse point cloud to perform selfcalibrated photogrammetry, wherein the system simultaneously solves for camera positions, intrinsic camera parameters, and three-dimensional point positions. In some embodiments, this approach is particularly advantageous for properties in remote or inaccessible locations where ground control points cannot be reliably established, and enables rapid deployment of the virtual inspection system without site-specific calibration.

[0368] In some embodiments, the alignment process uses an Iterative Closest Point (ICP) technique to iteratively minimize the distance between the estimated perimeter of the building structure derived from the sparse point cloud and the geospatially calibrated reference polygon provided by third-party data sources or machine learning-based structure footprint detection. In some embodiments, the technique applies successive rigid transformationsAttorney Docket No. FWIG-003W001 (rotation, translation, and uniform scaling) to the 3D point cloud until convergence, with convergence determined by monitoring the mean absolute distance between corresponding point pairs. In some embodiments, the scaling transformation accounts for potential systematic errors in the Structure-from-Motion reconstruction, such as those caused by long focal length variations across the image sequence or systematic scale drift in sequential feature tracking. In some embodiments, the alignment process validates the final transformation by confirming that the scaled estimated perimeter falls within a threshold distance of the reference polygon, indicating successful alignment.

[0369] In some embodiments, the property app displays aerial imagery overlaid with structural outlines and tree indicators (such as is described herein), allowing trained operators to manually adjust positions and attributes. In some embodiments, the system may employ a multi-model architecture where independent trained computer vision models are deployed for each structural feature category (windows, doors, vents, material transitions, roof features). In some embodiments, each model may be trained using positive examples of the specific feature type and negative examples of similar-appearing features and background objects. In some embodiments, this category-specific training approach enables each model to learn distinctive characteristics of its respective feature type. For example, a window detection model may learn to distinguish windows from skylights, reflective surfaces, and openings, while a door detection model may learn to distinguish doors from windows and deck railings. In some embodiments, the system may apply consensus-based detection, wherein a structural feature is confirmed as detected only if multiple independent models (or multiple passes of the same model with different confidence thresholds) identify it. In some embodiments, this ensemble approach may reduce false positive detections compared to single-model detection.

[0370] In some embodiments, the system generates bounding regions from detected structural features by identifying the minimum-area bounding polygon that encloses all points in a feature’s colored point cluster in the three-dimensional reconstruction. In some embodiments, for features detected in the 3D sparse point cloud, the system computes an oriented bounding box (OBB) by principal component analysis of the point cluster, computing the eigenvectors of the covariance matrix to determine the optimal axis-aligned orientation in 3D space. In some embodiments, the system projects this 3D-oriented bounding box onto the image plane and stores both the image-space coordinates (for display overlay) and the corresponding geospatial coordinates (for spatial reference and mitigation action tracking). In some embodiments, when the volumetrically rendered 3D reconstruction isAttorney Docket No. FWIG-003W001 scaled or translated during geospatial alignment, the system applies the corresponding scaling and translation transforms to the bounding region’s geospatial coordinates, ensuring dimensional measurements derived from bounding regions remain accurate relative to the aligned geospatial footprint.

[0371] In some embodiments, the auto-detection review engine enables users to efficiently process automatically detected tagged objects while providing tools to correct detection errors. In some embodiments, the system displays each automatically detected object along with a confidence indicator. In some embodiments, objects are initially sorted by confidence in descending order, so high-confidence detections are reviewed first. In some embodiments, for each automatically detected object, the user can: (1) accept the detection with or without annotation modifications, (2) adjust the bounding region to better fit the detected feature, (3) remove the detection if it represents a false positive, or (4) flag the detection for secondary review if ambiguous. In some embodiments, user corrections are logged and periodically used to retrain the detection models using techniques such as active learning, where corrections to low-confidence detections are prioritized for retraining to improve model performance on uncertain cases. In some embodiments, retraining is performed on a server infrastructure, and updated model weights are periodically deployed to the virtual inspection apparatus.

[0372] In some embodiments, when corrections are made in the property app 4330, a diff generator 4334 generates differential data (“diffs”) 4336 representing only the changes to be applied to the property features file, rather than regenerating the entire file. In some embodiments, the diff generator creates differential change records in JSON format that capture only the modifications made. In some embodiments, each diff record includes a path specifying the location within the property features hierarchy, an operation type, and the new value for add or replace operations. In some embodiments, when diffs are applied, the system processes them in a deterministic order to ensure reproducible results. In some embodiments, the diffs 4336 are stored as JSON and applied to the initial property feature file to produce an updated property feature file 4340. In some embodiments, the updated property feature file 4340 may be returned to the property app 4330 for additional corrections. In some embodiments, the system detects conflicts when two diffs target the same feature and values are logically inconsistent (for example, two diffs that simultaneously modify percent start and percent end of a span such that percent start exceeds percent end). In some embodiments, conflict detection triggers generation of an error report, and the conflicting diffs are rejected;Attorney Docket No. FWIG-003W001 the operator must manually resolve the conflict and resubmit. In some embodiments, this approach ensures that the property features file remains in a valid state after each diff application.

[0373] In some embodiments, a state machine workflow tracks the property feature through multiple states during the assessment lifecycle. In some embodiments, the state machine may define states such as “ready for repair” (e.g., initial file is available for human correction), “processing” (e.g., human corrections via diffs have been submitted and are being integrated), and “ready for review” (e.g., updated property features file is complete and awaiting supervisory review). In some embodiments, transitions between states are enforced by the system: for example, in the “ready for repair” state may only transition to “processing” upon receipt of a valid diff file that applies cleanly to the current property features file. In some embodiments, if a diff cannot be applied (e.g., due to intermediate changes or schema incompatibility), the state reverts to “ready for repair” and an error notification is generated. In some embodiments, this state machine ensures that property features files are never processed by PIM in an inconsistent or intermediate state.

[0374] In some embodiments, the updated property feature file 4340 provides input to a PIM 4350 that calculates wildfire ignition risk. In some embodiments, the PIM 4350 generates a property ignition report 4352.

[0375] In some embodiments, property features components used by PIM 4360 include structures (including wall spans with materials), trees (including trim information), elevation data, roof data, and material designations including the consider burnable flag.

[0376] In some embodiments, supplementary data 4370 stored in the property features file (but in some embodiments not directly processed by PIM) includes other shapes (such as decks and fences), observations, notes, and ground images. In some embodiments, supplementary data may be used to generate “violation” notes for homeowners or to document mitigation efforts over time.

[0377] In some embodiments, the geospatial tracking system uses the mobile device’s GPS receiver to record the service provider’s location while mitigation actions are performed. In some embodiments, the system establishes a geofence around each tagged object. In some embodiments, when the service provider’s GPS position enters the geofence, the system records the entry timestamp and monitors the provider’s location for the duration of the mitigation action. In some embodiments, the system can automatically confirm completionAttorney Docket No. FWIG-003W001 when the provider remains within the geofence for a configurable duration. In some embodiments, the system may account for GPS positioning error by expanding the geofence radius accordingly. In some embodiments, after all tagged objects are processed, the system may generate a compliance report that includes a map showing geofences and the service provider’s trajectory, timestamps for geofence entries, and a list of completed actions and any unvisited objects. In some embodiments, this report serves as automated verification that mitigation actions were performed.

[0378] Fig. 44 is a diagram illustrating geospatially located wall spans according to some embodiments.

[0379] Section A shows a structure shape representation 4400 according to some embodiments. In some embodiments, a structure footprint 4402 is defined as a polygon using geospatial coordinates, where each vertex of the polygon has associated latitude and longitude values. In some embodiments, a JSON representation 4404 stores the shape in WKT format as a POLYGON with coordinate pairs.

[0380] Section B shows a wall spans percentage-based representation 4410 according to some embodiments. In some embodiments, a wall linestring 4412 connects a start point (e.g., latitude and longitude of the start) to an end point (e.g., latitude and longitude of the end) and represents 0% to 100% of the wall length. In some embodiments, a window span 4414 may be positioned using percent start and percent end values (e.g., 10% to 25%). In some embodiments, a door span 4416 may be similarly positioned (e.g., 35% to 55%). In some embodiments, a material span 4418 designating a different wall material (e.g., brick) may be positioned (e.g., 70% to 88%).

[0381] Section C shows geospatial derivation 4420 according to some embodiments. In some embodiments, the geospatial coordinates of any point along the wall can be calculated from the percentage values using interpolation formulas 4422 (where span lat is latitude span, lat start is latitude start, pct is percent, lat end is latitude end, span ion is longitude span, lon start is longitude start, and lon end is longitude end): span lat = lat start + (pct / 100) x (lat end - lat start) and span ion = lon start + (pct / 100) x (lon end - lon start). In some embodiments, the percentage representation prevents orphan spans from coordinate drift during shape updates, as the inventors have recognized and appreciated that spans stored as independent linestrings could become detached from walls when structure shapes are modified.Attorney Docket No. FWIG-003W001

[0382] Section D shows coordinate systems 4430 according to some embodiments. In some embodiments, storage uses WGS84 / EPSG:4326 (spherical latitude / longitude coordinates). In some embodiments, editing uses EPSG:3857 (projected 2D coordinates for visualization) 4432. In some embodiments, reverse projection is performed when saving changes back to the property features file. In some embodiments, the system employs a dual-coordinate-system approach to optimize both data storage and user interaction. In some embodiments, the property features data structure stores all geometric data in WGS84 / EPSG:4326 (geographic coordinate reference system based on the World Geodetic System 1984), which uses latitude and longitude in degrees and is the international standard for geospatial data interchange and integration with external GIS systems and aerial imagery providers. In some embodiments, during visual inspection and editing, the system transforms the data to EPSG:3857 (Web Mercator projection), which projects the geographic coordinates onto a planar surface suitable for two-dimensional display and user interaction on desktop and mobile devices. In some embodiments, this projection is mathematically convenient for touch-based interactions and distance calculations in pixel space. In some embodiments, when the user completes edits, the system applies the inverse transformation to convert the edited coordinates back to WGS84ZEPSG:4326. In some embodiments, the transformation matrices for both forward and reverse projections are computed based on the centroid of the property and the zoom level, minimizing distortion for the specific property area being inspected.

[0383] The inventors have recognized and appreciated that materials and features may be geospatially located in the property features data structure, rather than merely being listed in an inspection report without spatial reference. In some embodiments, this geospatial association enables precise identification of feature locations for service provider guidance and accurate ignition modeling.Example Property Features Other Shapes Data Model

[0384] Fig. 45 is a block diagram of an other shapes data model according to some embodiments. In some embodiments, an other shape object 4500 represents miscellaneous facts and features that may not be formally included in other sections of the property features data structure. In some embodiments, the other shape object 4500 includes an internal identifier, a type specified by a type enumeration, a shape (polygon or point in geospatial coordinates), and an additional data field for schema-less additional attributes.Attorney Docket No. FWIG-003W001

[0385] In some embodiments, standalone features 4510 include types such as deck, pergola, garage, vegetative fuel sources, other fuel sources, fire hydrant, site access, spray plan, retardant container, community infrastructure, inspector assessment, and any other suitable feature type.

[0386] In some embodiments, attached features 4520 include types such as attached deck, attached fence, attached pergola, and any other suitable attached feature type. In some embodiments, attached features are linked to the structure and their positions are defined relative to the structure. An additional data examples section 4522 shows example additional data such as deck height values, hazard flags, and inspection notes.

[0387] In some embodiments, structure view data 4530 includes types for capturing directional data about the structure such as structure front data, structure front ignition zone, structure front view out, structure rear data, structure rear ignition zone, structure rear view out, structure left side data, structure left ignition zone, structure left view out, structure right side data, structure right ignition zone, and structure right view out.

[0388] In some embodiments, the other shapes data model is used for documentation of items that PIM does not directly model, such as access routes for emergency responders, fire retardant spray plans, and retardant container locations. In some embodiments, examples of such items include access routes, spray plans, and retardant containers. In some embodiments, the other shapes data model stores geospatial locations for service provider guidance. In some embodiments, decks and fences stored in other shapes may include height information (e.g., whether a deck is under 5 feet off the ground). In some embodiments, other shapes may be extended to capture data for additional perils beyond wildfire, such as flood or hurricane risk factors.

[0389] In some embodiments, extensibility is provided through the additional data field, which stores schema-less additional attributes not yet formalized in the property features schema. In some embodiments, this allows new data types to be captured without requiring immediate schema modifications, with a path to promoting frequently-used additional data fields to formal schema fields in future versions. In some embodiments, when a candidate field is validated by domain experts as critical to property assessment, it may be promoted to a formal schema field in a future schema version. In some embodiments, this rolling promotion process ensures the schema evolves to incorporate new knowledge and business requirements without requiring backward-incompatible changes. In some embodiments, whenAttorney Docket No. FWIG-003W001 a field is promoted, existing properties with that data in additional data may be automatically migrated by moving the values to the new formal field during a property features file update.Example Virtual Inspection Platform Architecture and Deployment

[0390] In some embodiments, the virtual inspection system is deployed as multiple application instantiations that share a common core codebase but present different user interfaces and feature sets based on deployment configuration. In some embodiments, the core property app provides authentication, property features file download / upload, and diff generation capabilities. In some embodiments, specific instantiations (such as Aerial Inspection and Virtual Inspection) add optional components (such as 2D overlay rendering for aerial imagery or 3D volumetrically-rendered visualization) and define their own inspection forms. In some embodiments, each instantiation has an associated configuration file that specifies: (1) which UI components are enabled, (2) the set of valid annotation categories and options for tagged objects, and (3) form field validation rules. For example, the Virtual Inspection app’s configuration enables the 3D model rendering component and includes additional form fields for “vantage point observation” (e.g., “visible from main entry”, “visible from northeast corner”) that are not available in the Aerial Inspection app. In some embodiments, this architecture enables rapid development of new inspection specializations without modifying the core platform. In some embodiments, when the core property app is updated (for example, to support a new data structure version), all instantiations may automatically inherit the update, ensuring consistency across the platform. In some embodiments, the system maintains version compatibility by implementing JSON schema validators that detect when a local property features file was generated by an incompatible version of the landscape pipeline or previous inspection app, and prompts the user to update before making modifications.Example System Benefits and Efficiency Improvements

[0391] In some embodiments, the virtual inspection system provides efficiency improvements over traditional on-site inspection for wildfire risk assessment. In some embodiments, a full inspection of a residential property using the disclosed system involves three steps: (1) image capture to acquire wrap-around video of the structure, (2) initial property features file generation via the landscape pipeline, and (3) virtual inspection and correction by trained operators. In some embodiments, the virtual inspection approach may reduce overall inspection time compared to traditional on-site inspection by eliminating travelAttorney Docket No. FWIG-003W001 time for each property and enabling operators in a central location to service multiple properties. In some embodiments, the three-dimensional rendering enables re-inspection and verification of completed mitigation actions without requiring the service provider to make a return visit to the property, further reducing operational costs.Example Computing Systems

[0392] Fig. 46 illustrates a networked computing system 4600 according to one or more embodiments. At least a portion of the networked computing system 4600 may be used to implement one or more methods and / or systems as described herein. The networked computing system 4600 may comprise a network 4602, at least one user 4604, at least one stationary computing device 4606, at least one mobile computing device 4608, and / or a server 4610, in any of various combinations as may be readily apprehended by one of ordinary skill in the art.

[0393] Network 4602 may be any one or more networks that allow elements of the present system to communicate with each other, as may be known in the art, for example wide area networks (WANs), LANs, and the like, that may be wired or wireless, and may include Bluetooth™, WiFi, and other approaches to local or remote communication.

[0394] There may be one or more stationary computing devices 4606 that may be one or more computing systems that may be used various users, such as homeowners (owners of building structures), appraisers, mitigation consultants or companies, monitoring consultants or companies, insurance companies or brokers and adjusters, structure data providers (such as images of structures, weather data, elevation data, material data, and the like).

[0395] One or more mobile computing devices 4608 may provide access to various functionality, similar to stationary computing device(s) 4606. In addition, mobile computing device(s) 4608 can include one or more digital cameras that can allow one or more users to capture digital images of the outside walls of a building structure, such as digital photographic images and / or video(s).

[0396] One or more stationary computing devices 4606, one or more mobile computing devices 4608, and / or one or more servers 4610 may be used to perform one or more methods as described herein - for example to generating a 3D sparse point cloud representation of a building structure, generating a 3D rendering of a building structure, performing a virtual inspection of a building structure, and / or estimating a wildfire risk for a building structure. AAttorney Docket No. FWIG-003W001 stationary computing device 4606, a mobile computing device 4608, or a server 4610 may be referred to simply as a computing device or a computer.

[0397] Fig. 47 is an example block diagram of a computing device 4700 that may incorporate one or more embodiments of the present disclosure. Fig. 47 is merely illustrative of a machine system to carry out aspects of the technical processes described herein, and does not limit the scope of the claims. One of ordinary skill in the art would recognize other variations, modifications, and alternatives. In one or more embodiments, the computing device 4700 typically includes a display and / or graphical user interface 4702, a data processing system 4720, a communication network interface 4712, input device(s) 4708, output device(s) 4706, and the like.

[0398] The data processing system 4720 may include one or more processors 4704 that communicate with a number of peripheral devices via a bus subsystem 4718. These peripheral devices may include the input device(s) 4708, the output device(s) 4706, the communication network interface 4712, and / or a storage subsystem, such as a volatile memory 4710 and a nonvolatile memory 4714.

[0399] The volatile memory 4710 and / or the nonvolatile memory 4714 may store computerexecutable instructions and thus forming logic 4722 that when applied to and executed by the processor(s) 4704 implement embodiments of the methods disclosed herein. The nonvolatile memory 4714 can store one or more trained ML models, one or more trained computer vision segmentation model, and / or one or more trained ML engines.

[0400] The input device(s) 4708 include devices and mechanisms for inputting information to the data processing system 4720. These may include a keyboard, a keypad, a touch screen incorporated into the monitor / di splay and / or graphical user interface 4702, audio input devices such as voice recognition systems, microphones, and other types of input devices. In various embodiments, the input device(s) 4708 may be embodied as a computer mouse, a trackball, a track pad, a joystick, wireless remote, drawing tablet, voice command system, eye tracking system, and the like. The input device(s) 4708 typically allow a user to select objects, icons, control areas, text and the like that appear on the monitor / display and / or graphical user interface 4702 via a command such as a click of a button or the like.

[0401] The output device(s) 4706 include devices and mechanisms for outputting information from the data processing system 4720. These may include the monitor / display and / orAttorney Docket No. FWIG-003W001 graphical user interface 4702, speakers, printers, infrared light emitting diodes (LEDs), and so on as understood in the art.

[0402] The communication network interface 4712 provides an interface to communication networks (e.g., one or more communication networks 4716) and devices external to the data processing system 4720. The communication network interface 4712 may serve as an interface for receiving data from and transmitting data to other systems. Embodiments of the communication network interface 4712 may include an Ethernet interface, a modem (telephone, satellite, cable, Integrated Services Digital Network (ISDN)), (asynchronous) digital subscriber line (DSL), FireWire, USB, a wireless communication interface such as BlueTooth or WiFi, a near-field communication wireless interface, a cellular interface, and / or another communication network interface.

[0403] The communication network interface 4712 may be coupled to the communication network(s) 4716 via an antenna, a cable, or the like. In one or more embodiments, the communication network interface 3512 may be physically integrated on a circuit board of the data processing system 4720, or in some cases may be implemented in software or firmware, such as “soft modems” or the like.

[0404] The computing device 4700 may include logic that enables communications over a network using protocols such as HTTP, TCP / IP, RTP / RTSP, IPX, UDP and the like.

[0405] The nonvolatile memory 4714 is an example of tangible (e.g., non-transitory) media configured to store computer-readable data and instructions to implement various embodiments of the processes described herein. The volatile memory 4710 is another example of tangible (e.g., non-transitory) media configured to store computer-readable data and instructions to implement various embodiments of the processes described herein. Other types of tangible media include removable memory (e.g., pluggable USB memory devices, mobile device SIM cards), optical storage media such as CD-ROMS, DVDs, semiconductor memories such as flash memories, non-transitory read-only-memories (ROMS), battery-backed volatile memories, networked storage devices, and the like. The volatile memory 4710 and / or the nonvolatile memory 4714 may be configured to store the basic programming and data constructs that provide the functionality of the disclosed methods and other embodiments thereof that fall within the scope of the present disclosure.

[0406] The logic 4722 that implements embodiments of the present disclosure may be stored in the volatile memory 4710 and / or in the nonvolatile memory 4714. The logic 4722 may beAttorney Docket No. FWIG-003W001 read from the volatile memory 4710 and / or non-volatile memory 4714 and executed by the processor(s) 4704. The volatile memory 4710 and / or the nonvolatile memory 4714 may also provide a repository for storing data used by the logic 4722.

[0407] The volatile memory 4710 and / or the nonvolatile memory 4714 may include a number of memories including a main random access memory (RAM) for storage of instructions and data during program execution and a read only memory (ROM) in which read-only non-transitory instructions are stored. The volatile memory 4710 and / or the nonvolatile memory 4714 may include a file storage subsystem providing persistent (nonvolatile) storage for program and data files. The volatile memory 4710 and / or the nonvolatile memory 4714 may include removable storage systems, such as removable flash memory.

[0408] The bus subsystem 4718 provides a mechanism for enabling the various components and subsystems of data processing system 4720 to communicate with each other. Although the communication network interface 4712 is depicted schematically as a single bus, some embodiments of the bus subsystem 4718 may utilize multiple distinct busses.

[0409] The computing device 4700 may be a device such as a smartphone, a desktop computer, a laptop computer, a rack-mounted computer system, a computer server, or a tablet computer device. The computing device 4700 may be implemented as a collection of multiple networked computing devices, for example in a distributed computing system. Further, the computing device 4700 may typically include operating system logic the types and nature of which are known in the art.The Disclosed and Claimed Systems, Apparatuses, and / or Methods Constitute Patent-Eligible Subject Matter

[0410] The principles underlying patent-eligible subject matter, at least in the United States, distinguish between: A) claimed subject matter that recites a law of nature, a natural phenomenon, or an “abstract idea;” and B) claimed subject matter that merely involves a law of nature, natural phenomenon, or an abstract idea (e.g., as a constituent element of a combination of elements that achieves one or more concrete technological improvements). In particular, claimed subject matter that recites a law of nature, a natural phenomenon, or an abstract idea requires further eligibility analysis during examination; in contrast, claimed subject matter that merely involves a law of nature, natural phenomenon, or an abstract idea is patent-eligible without further analysis during examination. This distinction is particularly relevant for inventive subject matter that, at least in part, relies on or pertains to one or moreAttorney Docket No. FWIG-003W001 of artificial intelligence (Al), machine learning (ML), data processing, computer vision, three-dimensional reconstruction, or geospatial data management.

[0411] Regarding the notion of an “abstract idea,” the United States Patent and Trademark Office (USPTO) has enumerated categories of abstract ideas as follows: 1) mathematical concepts (e.g., mathematical relationships, mathematical formulas or equations, mathematical calculations); 2) certain methods of organizing human activity, which include “fundamental economic principles or practices” (e.g., hedging, insurance, financial transactions), “commercial or legal interactions” (e.g., advertising, marketing, sales agreements, contract formation), and “managing personal behavior and relationships” (e.g., following rules, social interactions); and 3) mental processes (e.g., concepts performed in the human mind, observation, evaluationjudgment, opinion, mental steps).

[0412] With respect to the second enumerated category above, some courts in the United States have referred to these “certain methods of organizing human activity” more generally as “abstract business methods.” Approaches for analyzing patent-eligible subject matter also recognize that not all subject matter with a commercial application constitutes an “abstract business method;” rather, inventive subject matter can of course have a commercial application and still be patent-eligible. It is only when subject matter involving economic principles, commercial interactions, or following rules, for example, is implemented using generic computing components - and without providing a genuine technical improvement -that it may be deemed an “abstract business method.”

[0413] The inventive systems, apparatuses, and methods disclosed and claimed herein relating to virtual wildfire inspection, three-dimensional building structure reconstruction, property features data management, and wildfire mitigation action tracking are not “abstract business methods” but instead clearly constitute inventive patent-eligible subject matter. This inventive patent-eligible subject matter may, in some instances, involve one or more so-called “abstract ideas” as effective tools to achieve technological improvements. However, claims directed to this inventive patent-eligible subject matter do not merely or exclusively recite a law of nature, natural phenomenon, or a so-called “abstract idea.” Instead, the inventive systems, apparatuses, and methods disclosed and claimed herein focus on specific multifaceted structural and / or functional technological solutions (e.g., generating geospatially calibrated three-dimensional renderings of building structures from wrap-around video, managing hierarchical property features data structures with percentage-based spatial representations, tracking differential changes to property assessment data, and guiding serviceAttorney Docket No. FWIG-003W001 providers through geospatially-located mitigation actions) and thus, prima facie, they are directed to patent-eligible subject matter.

[0414] To the extent any of the disclosed and claimed inventive subject matter might be characterized in part as reciting one or more abstract ideas, the analysis set forth below demonstrates that any such abstract idea is nonetheless integrated into one or more practical applications that provide concrete technological improvements. In particular, any claimed subject matter that might be characterized in part as reciting one or more abstract ideas nonetheless explicitly integrates the abstract idea(s) into a patently practical application, namely, providing important technological solutions to address one or more of the overtly harrowing problems posed by wildfires (e.g., widespread structural and / or environmental damage, efficient and accurate assessment of properties vulnerable to such damage, and effective guidance for mitigation actions to reduce wildfire risk).Example Technological Solutions to Technological Problems in Virtual Wildfire Inspection and Property Features Data Management

[0415] In developing the inventive systems, apparatuses, and methods disclosed and claimed herein, the inventors recognized and appreciated multiple technological problems with conventional techniques for wildfire vulnerability assessment, property inspection, and wildfire mitigation tracking. As introduced above and discussed in further detail below, the inventors have addressed and overcome these technological problems with innovative and specific technological solutions to effectively realize the various technical features described herein.

[0416] More specifically, examples of one or more problems addressed by the specific technological solutions provided by the inventive systems, apparatuses, and methods disclosed and claimed herein include, but are not limited to:

[0417] 1. How to enable remote virtual inspection of a building structure’s exterior without requiring a skilled technician to travel to the property, while still capturing sufficient detail to identify structural features relevant to wildfire ignition risk;

[0418] 2 How to generate a geospatially calibrated three-dimensional rendering of a building structure from video captured by a person walking around the structure with a mobile device, and how to align that rendering with authoritative geospatial reference data;Attorney Docket No. FWIG-003W001

[0419] 3 How to represent structural features (such as windows, doors, vents, and material variations) on a building structure’s walls in a data structure that automatically maintains correct relative positions when the underlying wall geometry is modified;

[0420] 4. How to track incremental corrections to property features data made by trained operators without regenerating entire data files, while ensuring data integrity through conflict detection and state machine workflow management;

[0421] 5. How to aggregate wildfire risk data from multiple authoritative sources and maintain version history for audit purposes;

[0422] 6. How to represent tree characteristics including trim information in a manner that enables calculation of defensible space requirements and quantification of potential risk reduction from trimming versus removal;

[0423] 7. How to guide a service provider who may be unfamiliar with a property to specific geospatially-located mitigation actions, and how to automatically verify completion of those actions using geospatial tracking; and

[0424] 8. How to enable property owners to model hypothetical mitigation scenarios and quantify the risk reduction benefit of proposed actions before committing resources.

[0425] With respect to the technological solutions to these respective technological problems, the inventive systems, apparatuses, and methods disclosed herein provide specific technical implementations that address each of the above problems.

[0426] Regarding the first and second technological problems, the disclosed systems generate a volumetrically-rendered, geospatially-calibrated three-dimensional reconstruction of a building structure from wrap-around video using Structure-from-Motion (SfM) techniques. The systems employ feature-based visual odometry to establish consistent camera pose estimation across the image sequence, enabling accurate three-dimensional reconstruction even when imaging conditions vary due to changes in lighting, reflectance properties of building materials, and weather conditions. The systems further employ an Iterative Closest Point (ICP) technique to align the three-dimensional reconstruction with a geospatially calibrated reference polygon, applying successive rigid transformations (rotation, translation, and uniform scaling) until convergence. This alignment process enables structural features identified in the reconstruction to be precisely located in geospatial coordinates, facilitating integration with property features data and enabling accurate guidance for service providers.Attorney Docket No. FWIG-003W001

[0427] Regarding the third technological problem, the disclosed systems employ a percentage-based wall span representation that stores structural feature positions as percentages of a wall linestring rather than as independent geospatial coordinates. The inventors have recognized and appreciated that this representation prevents orphaned spans that would occur if spans were stored as independent linestrings, because the span positions automatically adjust when the wall shape is modified due to corrected structural measurements or updated GIS data. The interpolation calculation for deriving geospatial coordinates from percentage values is performed with double-precision floating-point arithmetic to ensure sub-meter accuracy in geospatial positioning even for very long walls in large properties.

[0428] Regarding the fourth technological problem, the disclosed systems employ a differential change tracking mechanism that generates differential change records (diffs) representing only the modifications made to the property features data structure, rather than regenerating the entire file. Each diff record includes a path specifying the location within the property features hierarchy, an operation type (add, replace, or remove), and the new value for add or replace operations. The systems further employ a state machine workflow that tracks the property feature through states such as “ready for repair,” “processing,” and “ready for review,” with enforced transitions that ensure property features files are never processed by a property ignition model in an inconsistent or intermediate state. Conflict detection identifies when two diffs target the same feature with logically inconsistent values, triggering error reports and rejection of conflicting diffs.

[0429] Regarding the fifth technological problem, the disclosed property features data structure stores risk data aggregated from multiple authoritative sources (such as public utilities commissions, state fire protection agencies, and federal agriculture agencies) rather than selecting a single value. Each risk value includes metadata such as a numeric risk value, risk label, source identifier, and load version indicating when the data was obtained. The systems maintain a version history of all risk data, enabling auditors to verify which risk values were in effect at the time a property assessment was conducted.

[0430] Regarding the sixth technological problem, the disclosed tree data model includes trim information comprising a maximum radius reduction (representing the maximum amount the canopy radius could be reduced by trimming), an actual trim height (representing the current trim height based on observed cut scars), a maximum trim height (representing the maximum height to which the tree could be trimmed), and a highest adjacent rooflineAttorney Docket No. FWIG-003W001 (representing the height of the highest roofline of adjacent structures). The tree data model further includes a breast height diameter measurement that may be used in conjunction with fire behavior research to calculate heat production during combustion. This structured representation enables a property ignition model to quantify the difference in wildfire risk between the current state and hypothetical trimmed or removed states.

[0431] Regarding the seventh technological problem, the disclosed systems establish a geofence around each tagged object representing a mitigation action. When a service provider’s GPS position enters the geofence, the systems record the entry timestamp and monitor the provider’s location for the duration of the mitigation action. The systems can automatically confirm completion when the provider remains within the geofence for a configurable duration. After all tagged objects are processed, the systems generate a compliance report including a map showing geofences and the service provider’s trajectory, timestamps for geofence entries, and a list of completed actions and any unvisited objects.

[0432] Regarding the eighth technological problem, the disclosed systems enable creation of a copy of the property features data structure that can be modified to reflect the expected state after proposed mitigation actions. The property ignition model can then be executed on both the current state and the modified state to compute the change in probability of structure destruction. The systems enable batching of multiple hypothetical scenarios and generating comparison reports showing the risk reduction for each scenario, enabling property owners and risk managers to make data-driven decisions about which mitigation actions provide the best cost-benefit ratio.

[0433] These technological solutions do not constitute so-called “abstract ideas.” Instead, these approaches provide specific technical implementations that also offer multiple improvements to computer functionality. In particular, the percentage-based wall span representation provides a storage-efficient mechanism that automatically maintains data integrity when underlying geometry changes. The differential change tracking mechanism reduces data transmission and storage requirements while enabling collaborative editing with conflict detection. The Structure-from-Motion and ICP alignment pipeline enables generation of geospatially accurate three-dimensional models from commodity mobile device cameras without requiring specialized surveying equipment or ground control points. The geofencebased mitigation verification provides automated confirmation of service completion without requiring manual reporting. Accordingly, these approaches represent improvements to the technical fields of computer vision, geospatial data management, and building informationAttorney Docket No. FWIG-003W001 modeling, particularly in connection with the practical application of wildfire vulnerability assessment and mitigation planning.Conclusion

[0434] In one or more embodiments, a PIM can be powered, at least in part, by ML engines or models that can be trained using previous wildfires including actual damage to building structures, the number and location of potential ignition sources (e.g., trees, brush, wood structures such as decks) on the properties, the weather in the days or weeks leading up to the wildfires, the building structure materials, and / or other inputs. Additionally or alternatively, a PIM can be physics based.

[0435] In one or more embodiments, a PIM can use as inputs remote sensors or devices that can be installed or located at a target property. The remote sensors / devices can provide realtime and / or updated data such as the temperature, relative humidity, wind speed, wind direction, barometric pressure, precipitation, and / or other data. The remote sensors / devices can be connected to a network, such as the internet, for example as loT sensors / devices. The PIM can use data from the remote sensors / devices as inputs in addition to other data such as weather forecasts, wildfire forecasts, and / or other data.

[0436] In one or more embodiments, a PIM can be used to generate real-time and / or updated alerts regarding the wildfire ignition risks for a target property. The alerts can be sent to the property owner / resident, an insurance agency, an insurance agent, or another person. The alerts can be sent over a wired and / or a wireless network.

[0437] In one or more embodiments, distributed ledgers, such as those using blockchain technology, can be used to secure storage and sharing of risk assessment data with regulatory bodies and / or insurers.

[0438] This disclosure should not be considered limited to the particular embodiments described above. Various modifications, equivalent processes, as well as numerous structures to which the present technology may be applicable, will be readily apparent to those skilled in the art to which embodiments disclosed herein are directed upon review of this disclosure. The above-described embodiments may be implemented in numerous ways. One or more aspects and embodiments involving the performance of processes or methods may utilize program instructions executable by a device (e.g., a computer, a processor, or other device) to perform, or control performance of, the processes or methods.Attorney Docket No. FWIG-003W001

[0439] In this respect, various inventive concepts may be embodied as a non- transitory computer readable storage medium (or multiple non-transitory computer readable storage media) (e.g., a computer memory of any suitable type including transitory or non-transitory digital storage units, circuit configurations in field programmable gate arrays (FPGAs) or other semiconductor devices, or other tangible computer storage medium) encoded with one or more programs that, when executed on one or more computers or other processors, perform methods that implement one or more of the various embodiments described above. When implemented in software (e.g., as an app), the software code may be executed on any suitable processor or collection of processors, whether provided in a single computer or distributed among multiple computers.

[0440] Further, it should be appreciated that a computer may be embodied in any of a number of forms, such as a rack-mounted computer, a desktop computer, a laptop computer, or a tablet computer (e.g., an iPad), as non-limiting examples. Additionally, a computer may be embedded in a device not generally regarded as a computer but with suitable processing capabilities, including a personal digital assistant (PDA), a smartphone or any other suitable portable or fixed electronic device.

[0441] Also, a computer may have one or more communication devices, which may be used to interconnect the computer to one or more other devices and / or systems, such as, for example, one or more networks in any suitable form, including a local area network or a wide area network, such as an enterprise network, an intelligent network, or the Internet. Such networks may be based on any suitable technology and may operate according to any suitable protocol and may include wireless networks or wired networks.

[0442] Also, a computer may have one or more input devices and / or one or more output devices. These devices can be used, among other things, to present a user interface. Examples of output devices that may be used to provide a user interface include printers or display screens for visual presentation of output and speakers or other sound generating devices for audible presentation of output. Examples of input devices that may be used for a user interface include keyboards, and pointing devices, such as mice, touch pads, and digitizing tablets. As another example, a computer may receive input information through speech recognition or in other audible formats.

[0443] The non-transitory computer readable medium or media may be transportable, such that the program or programs stored thereon may be loaded onto one or more differentAttorney Docket No. FWIG-003W001 computers or other processors to implement various one or more of the aspects described above. In some embodiments, computer readable media may be non- transitory media.

[0444] The terms "program," “app,” and "software" are used herein in a generic sense to refer to any type of computer code or set of computer-executable instructions that may be employed to program a computer or other processor to implement various aspects as described above. Additionally, it should be appreciated that, according to one aspect, one or more computer programs that when executed perform methods of the present application need not reside on a single computer or processor but may be distributed in a modular fashion among a number of different computers or processors to implement various aspects of the present application.

[0445] Computer-executable instructions may be in many forms, such as program modules, executed by one or more computers or other devices. Generally, program modules include routines, programs, objects, components, data structures, etc. that performs particular tasks or implement particular abstract data types. The functionality of the program modules may be combined or distributed as desired in various embodiments.

[0446] Also, data structures may be stored in computer-readable media in any suitable form. For simplicity of illustration, data structures may be shown to have fields that are related through location in the data structure. Such relationships may likewise be achieved by assigning storage for the fields with locations in a computer-readable medium that convey relationship between the fields. However, any suitable mechanism may be used to establish a relationship between information in fields of a data structure, including through the use of pointers, tags or other mechanisms that establish relationship between data elements.

[0447] Thus, the present disclosure includes new and novel improvements to existing methods and technologies, which were not previously known nor implemented to achieve the useful results described above. Users of the present method and system will reap tangible benefits from the functions now made possible on account of the specific modifications described herein causing the effects in the system and its outputs to its users. It is expected that significantly improved operations can be achieved upon implementation of some embodiments disclosed herein.

[0448] Also, as described, some aspects may be embodied as one or more methods. The acts performed as part of the method may be ordered in any suitable way. Accordingly, embodiments may be constructed in which acts are performed in an order different thanAttorney Docket No. FWIG-003W001 illustrated, which may include performing some acts simultaneously, even though shown as sequential acts in illustrative embodiments.

Claims

Attorney Docket No. FWIG-003W001CLAIMS1. A computer-implemented method for tracking wildfire mitigation actions, comprising: displaying, on a display screen of a mobile computing device associated with a user, a displayed portion of a volumetrically rendered three-dimensional reconstruction of outside walls of a building structure;displaying, on the display screen, one or more tagged objects, each tagged object representing a respective physical object and / or a respective physical area where one or more assigned wildfire mitigation actions is / are to be performed;receiving, with the mobile computing device, a selection of one of the one or more tagged objects to form a selected tagged object;providing, with the mobile computing device, a physical location associated with the selected tagged object;providing, with the mobile computing device, the one or more assigned wildfire mitigation actions associated with the selected tagged object;geospatially tracking the mobile computing device while the user performs the one or more assigned wildfire mitigation actions associated with the selected tagged object;requesting, with the mobile computing device, a confirmation that the user performed the one or more assigned wildfire mitigation actions associated with the selected tagged object;receiving, with the mobile computing device, the confirmation that the user performed the one or more assigned wildfire mitigation actions associated with the selected tagged object;updating, in response to the confirmation and with the mobile computing device, a status of the selected tagged object; andrequesting, in response to the confirmation and with the mobile computing device, that the user capture digital images of all the outside walls of the building structure including the physical location associated with the selected tagged object.

2. The method of claim 1, further comprising providing, with the mobile computing device, a geolocation associated with the selected tagged object.

3. The method of claim 2, further comprising:determining a current geolocation of the mobile computing device; andAttorney Docket No. FWIG-003W001 providing directions, with the mobile computing device, to direct the user from the current geolocation to the geolocation associated with the selected tagged object.

4. The method of claim 1, further comprising receiving, with the mobile computing device, one or more user inputs representing the confirmation that the user performed the one or more assigned wildfire mitigation actions associated with the selected tagged object.

5. The method of claim 4, wherein the step of requesting, with the mobile computing device, the confirmation that the user performed the one or more assigned wildfire mitigation actions associated with the selected tagged object includes prompting the user, with the mobile computing device, to capture one or more digital images of the respective physical object and / or the respective physical area associated with the selected tagged object, the one or more digital images representing a completion of the one or more assigned wildfire mitigation actions associated with the selected tagged object.

6. The method of claim 5, further comprising:capturing the one or more digital images of the respective physical object and / or the respective physical area associated with the selected tagged object, the one or more digital images representing the completion of the one or more assigned wildfire mitigation actions associated with the selected tagged object; andsending the one or more digital images from the mobile computing device to a second computer in communication with the mobile computing device.

7. The method of claim 5, wherein the step of updating, in response to the confirmation and with the mobile computing device, the status of the selected tagged object includes changing a visual appearance of the selected tagged object on the display screen.

8. A system for performing a virtual inspection of a building structure on a property to determine a risk of loss due to wildfire, the building structure having outside walls, the system comprising:a data store that includes image data that represent a geospatially registered three-dimensional rendering of the outside walls of the building structure and property features data that represent tagged objects, the property features data including indexed image positions and corresponding indexed geospatial positions of the tagged objects;Attorney Docket No. FWIG-003W001 a virtual inspection engine in communication with the data store, the virtual inspection engine configured to:display a displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the building structure according to a position and field-of-view (FOV) of a virtual camera relative to the geospatially registered three- dimensional rendering of the outside walls of the building structure, the displayed portion including one or more of the tagged objects,overlay one or more two-dimensional bounding regions over the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the building structure, each of the one or more two-dimensional bounding regions having a respective indexed image position corresponding to a respective one or more of the tagged objects, the respective indexed image position included in the property features data,receive a user input that represents a modification to one or more properties of a target two-dimensional bounding region for a target tagged object, andsend an update to the data store to update the property features data according to the modification to the one or more properties of the target two-dimensional bounding region.

9. The system of claim 8, wherein the modification to the one or more properties of the target two-dimensional bounding region includes a position, a size, and / or a shape of the target two-dimensional bounding region.

10. The system of claim 8, wherein the update to the property features data includes an update to the respective indexed image position of the target tagged object and a corresponding update to an indexed geospatial position of the target tagged object.

11. The system of claim 8, wherein the update to the property features data includes an update to a respective indexed image size of the target tagged object and a corresponding update to an indexed geospatial size of the target tagged object, the indexed geospatial size representing a virtual measurement of dimensions of the target tagged object.

12. The system of claim 8 wherein:the user input is a first user input, andAttorney Docket No. FWIG-003W001 the virtual inspection engine is further configured to:receive a second user input that represents a tag of an untagged object represented in the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the building structure, the tag of the untagged object resulting in a new tagged object,define a new two-dimensional bounding region for the new tagged object, the new two-dimensional bounding region having a new indexed image position, overlay the new two-dimensional bounding region over the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the building structure,wherein the update to the data store includes an addition of the new tagged object to the property features data.

13. The system of claim 12, wherein:the virtual inspection engine is further configured to receive annotation data for the new tagged object, andthe update to the data store includes the annotation data for the new tagged object.

14. The system of claim 10, wherein the system further comprises an image processor having an image recognition engine configured to:automatically detect structural features of the building structure in the geospatially registered three-dimensional rendering of the outside walls of the building structure, and map the structural features automatically detected by the image recognition engine to an image space of the geospatially registered three-dimensional rendering of the outside walls of the building structure,wherein the image processor is configured to store the structural features automatically detected by the image recognition engine as at least some of the tagged objects in the property features data stored in the data store.

15. The system of claim 8, wherein the virtual inspection engine includes:a rendering engine configured to produce output data that represent the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the building structure, andAttorney Docket No. FWIG-003W001 a tagged object rendering engine configured to generate each of the one or more two-dimensional bounding regions according to the respective indexed image position corresponding to the respective one or more of the tagged objects.

16. The system of claim 15, wherein:the user input is a first user input,the virtual inspection engine further includes a user input detector configured to detect a second user input that represents a change in the position and / or the FOV of the virtual camera,the rendering engine is configured to update the output data that represent the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the building structure according to the change in the position and / or the FOV of the virtual camera, andthe tagged object rendering engine is configured to update the one or more two-dimensional bounding regions generated according to the change in the position and / or the FOV of the virtual camera.

17. The system of claim 15, wherein:the image data in the data store further represents an overhead rendering of a target property that includes the building structure,the one or more of the tagged objects displayed in the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the building structure is / are one or more first tagged objects,the one or more bounding regions is / are one or more first bounding regions, the virtual inspection engine is further configured to overlay the overhead rendering of the target property over the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the building structure, the overhead rendering including one or more second bounding regions for one or more second tagged objects represented in the overhead rendering.

18. The system of claim 17, wherein:the one or more first tagged objects is / are one or more structural features of the building structure, andthe one or more second tagged objects is / are one or more trees in the target property.Attorney Docket No. FWIG-003W00119. A virtual wildfire inspection apparatus comprising:a display screen;one or more microprocessors in communication with the display screen;non-transitory computer memory in communication with the one or more microprocessors, the non-transitory computer memory storing property features data for a target building structure on a target property, the property features data including:image data representing a volumetrically rendered geospatially calibrated 3D reconstruction of the outside walls of the target building structure, andtagged object data representing tagged objects on the outside walls of the target building structure, the tagged object data including a respective indexed position of each tagged object and annotation data for each tagged object,the non-transitory computer memory further storing computer-readable instructions that, when executed by the one or more microprocessors, cause the one or more microprocessors to run a virtual inspection engine configured to:display, using the image data, a displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the target building structure according to a position and field-of-view (FOV) of a virtual camera relative to the geospatially registered three-dimensional rendering of the outside walls of the target building structure, the displayed portion including one or more of the tagged objects,overlay one or more two-dimensional bounding regions over the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the target building structure, each of the one or more two-dimensional bounding regions having a respective indexed image position corresponding to a respective one or more of the tagged objects;receive a user input that represents a modification to one or more properties of a target two-dimensional bounding region for a target tagged object, and generate an update to the property features data stored in the non-transitory computer memory according to the modification to the one or more properties of the target two-dimensional bounding region.Attorney Docket No. FWIG-003W001 20. The apparatus of claim 19, wherein the modification to the one or more properties of the target two-dimensional bounding region includes a position, a size, and / or a shape of the target two-dimensional bounding region.

21. The apparatus of claim 20, wherein the update to the property features data includes an update to the respective indexed image position of the target tagged object and a corresponding update to an indexed geospatial position of the target tagged object.

22. The apparatus of claim 20, wherein the update to the property features data includes an update to a respective indexed image size of the target tagged object and a corresponding update to an indexed geospatial size of the target tagged object, the indexed geospatial size representing a virtual measurement of dimensions of the target tagged object.

23. The apparatus of claim 19, wherein:the user input is a first user input, andthe virtual inspection engine is further configured to:receive a second user input that represents a tag of an untagged object represented in the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the target building structure, the tag of the untagged object resulting in a new tagged object,define a new two-dimensional bounding region for the new tagged object, the new two-dimensional bounding region having a new indexed image position, overlay the new two-dimensional bounding region over the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the target building structure,wherein the update to the property features data includes an addition of the new tagged object to the property features data.

24. The apparatus of claim 23, wherein:the virtual inspection engine is further configured to receive annotation data for the new tagged object, andthe update to the property features data includes the annotation data for the new tagged object.Attorney Docket No. FWIG-003W001 25. The apparatus of claim 19, wherein:the image data further represents an overhead rendering of the target property, the one or more of the tagged objects displayed in the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the target building structure is / are one or more first tagged objects,the one or more bounding regions is / are one or more first bounding regions, the virtual inspection engine is further configured to overlay the overhead rendering of the target property and the displayed portion of the geospatially registered three-dimensional rendering of the outside walls of the target building structure, the overhead rendering including one or more second bounding regions for one or more second tagged objects represented in the overhead rendering.

26. A system for determining a risk of loss to a building structure on a property due to wildfire, the building structure having outside walls, the system comprising:a first data store that receives and stores image data from one or more image sources, the image data including a collective representation of at least a portion of the property and all of the outside walls of the building structure including structural features on and / or in the outside walls and a geospatially calibrated overhead view of the property including the building structure;an image processing engine in communication with the first data store, the image processing engine including one or more image processors, the image processing engine configured to:detect the outside walls of the building structure and the structural features on and / or in the outside walls;generate encoded structural descriptors that include one or more properties of the outside walls of the building structure, one or more properties of the structural features on and / or in the outside walls, and indexed structural locations of the outside walls and the structural features;detect one or more fuel sources on the at least a portion of the property; generate encoded fuel descriptors that include one or more properties of each of the one or more fuel sources and one or more indexed fuel locations of the one or more fuel sources, respectively;Attorney Docket No. FWIG-003W001 a second data store that receives and stores the encoded structural descriptors and the encoded fuel descriptors, the second data store in communication with the image processing engine;a wildfire risk modeling engine in communication with the second data store, the wildfire risk modeling engine including one or more risk processors configured to execute computer-readable instructions corresponding to a wildfire loss risk model, the wildfire risk modeling engine receiving the encoded structural descriptors and the encoded fuel descriptors as inputs and producing at least one output;an annotation engine in communication with the wildfire risk modeling engine, the first data store, and the second data store, the annotation engine including one or more annotation processors, the annotation engine configured to:display, using the image data, a geospatially registered rendering of the outside walls of the building structure and objects on the at least a portion of the property, the objects including uncategorized objects and the one or more fuel sources;generate, using the encoded structural descriptors, structural bounding regions, each structural bounding region corresponding to a respective indexed structural location of a respective structural feature;generate, using the encoded fuel source descriptors, one or more fuel source bounding regions, each fuel source bounding region corresponding to a respective indexed fuel location of a respective fuel source;overlay the structural bounding regions and the one or more fuel source bounding regions on the geospatially registered rendering, the structural bounding regions and the one or more fuel source bounding regions configured to be adjustable in response to user input;send an update to the second data store to modify at least one of the indexed structural locations and / or the one or more indexed fuel locations based on the user input,wherein the wildfire risk modeling engine is configured to update the at least one output in response to the update to the second data store.

27. The system of claim 26, wherein the geospatially registered rendering of the outside walls of the building structure comprises a geospatially registered three-dimensional rendering of the outside walls of the building structure.Attorney Docket No. FWIG-003W00128. The system of claim 26, wherein the geospatially registered rendering of the outside walls of the building structure comprises an overhead rendering of the property.

29. The system of claim 26, wherein:the user input is a first user input,the annotation engine is further configured to receive annotation data, in response to a second user input, the annotation data that include annotations that update and / or supplement the one or more properties of the structural features on and / or in the outside walls of the building structure, wherein the update to the second data store includes the annotation data.

30. The system of claim 26, wherein:the user input is a first user input,the annotation engine is further configured to:receive a second user input that defines a new structural bounding region and annotated structural descriptors for a first uncategorized object of the uncategorized objects,determine a new indexed structural location for the uncategorized object, the first indexed structural location corresponding to the new structural bounding region, andsend an update to the second data store to add the new indexed structural location to the encoded structural descriptors to produce updated structural descriptors,wherein the wildfire risk modeling engine is configured to update the at least one output based on the updated encoded structural descriptors.

31. A computer-implemented method for managing property assessment data for wildfire risk assessment, the method comprising:A) storing, in a non-transitory computer memory, a property features data structure representing a property including one or more building structures, the property features data structure including:structures data containing one or more structure objects, each structure object representing a respective building structure of the one or more building structures and including a structure shape defined as a polygon in geospatial coordinates, a structureAttorney Docket No. FWIG-003W001 type field indicating whether the respective building structure is a main structure or a fuel structure, and walls data containing one or more wall objects;trees data containing one or more tree objects, each tree object representing a respective tree on or near the property and including a tree shape defined as a polygon representing a canopy footprint of the respective tree, a height value, and trim information; andan elevation section containing terrain elevation data including geographic bounds and a plurality of elevation values representing elevation above sea level; B) receiving, from a virtual inspection application executed by one or more processors, a modification to one or more values in the property features data structure, the modification representing a correction identified by an operator;C) generating, with the one or more processors, a differential change record representing the modification, the differential change record including:a path specifying a hierarchical location within the property features data structure where the modification is to be applied,an operation type selected from add, replace, or remove, anda new value for the modification when the operation type is add or replace; D) storing the differential change record in the non-transitory computer memory separately from the property features data structure; andE) applying the differential change record to the property features data structure to produce an updated property features data structure.

32. The method of claim 31, wherein the property features data structure further includes an info section containing property metadata, the property metadata including:an address of the property,a center of analysis represented as latitude and longitude coordinates, andrisk data aggregated from a plurality of authoritative sources.

33. The method of claim 32, wherein the risk data includes a plurality of risk values from two or more sources selected from:a public utilities commission, a state fire protection agency, and a federal agriculture agency, and wherein each risk value of the plurality of risk values includes:a numeric risk value,a risk label,Attorney Docket No. FWIG-003W001 a source identifier identifying a source of the risk value, anda load version indicating a date or time when the risk value was obtained.

34. The method of claim 31, wherein the property features data structure is stored as a JSON file using a WGS84 / EPSG:4326 coordinate reference system for the geospatial coordinates.

35. The method of claim 31, further comprising:F) tracking a workflow state of the property features data structure using a state machine, the state machine defining a plurality of states including a ready-for-repair state, a processing state, and a ready -for-review state;G) transitioning the workflow state from the ready-for-repair state to the processing state in response to receiving the differential change record; andH) transitioning the workflow state from the processing state to the ready-for-review state in response to successfully applying the differential change record to produce the updated property features data structure.

36. The method of claim 35, wherein when applying the differential change record fails due to an intermediate change to the property features data structure or a schema incompatibility, the workflow state reverts to the ready-for-repair state and an error notification is generated.

37. The method of claim 31, further comprising:F) receiving a plurality of differential change records;G) detecting a conflict when two differential change records of the plurality of differential change records target a same element within the property features data structure with logically inconsistent values; andH) in response to detecting the conflict, generating an error report and rejecting the two differential change records.

38. The method of claim 31, wherein:the property features data structure further includes other shapes data containing one or more other shape objects, each other shape object representing a miscellaneous property feature not formally categorized in the structures data or the trees data, and wherein the miscellaneous property feature is a deck, a fence, a fuel source, a fire hydrant, a site access route, a spray plan, or a retardant container location, andAttorney Docket No. FWIG-003W001 the method further comprises:F) creating a modified copy of the property features data structure, the modified copy reflecting an expected state of the property after one or more proposed mitigation actions;G) executing a property ignition model on the property features data structure to calculate a first probability of destruction;H) executing the property ignition model on the modified copy to calculate a second probability of destruction;I) computing a risk reduction value as a difference between the first probability of destruction and the second probability of destruction, the risk reduction value quantifying a benefit of the one or more proposed mitigation actions;J) creating a plurality of modified copies of the property features data structure, each modified copy of the plurality of modified copies reflecting a different set of proposed mitigation actions;K) executing the property ignition model on each modified copy of the plurality of modified copies to calculate a respective probability of destruction for each modified copy; andL) generating a comparison report showing a respective risk reduction value for each modified copy, enabling selection of proposed mitigation actions based on risk reduction benefit.

39. A computer-implemented method for representing structural features along a wall of a building structure, the method comprising:A) storing, in a non-transitory computer memory, a wall object representing a wall segment of the building structure, the wall object including a wall linestring defining the wall segment in geospatial coordinates, the wall linestring having a start point and an end point;B) storing, in the non-transitory computer memory and in association with the wall object, one or more wall span objects, each wall span object of the one or more wall span objects representing a respective structural feature along the wall segment and including:a span type indicating a type of the respective structural feature, a percentage start value representing a starting position of the respective structural feature as a percentage of a length of the wall linestring, anda percentage end value representing an ending position of the respective structural feature as a percentage of the length of the wall linestring;Attorney Docket No. FWIG-003W001 C) receiving, at one or more processors, a modification to the wall linestring that changes at least one of the start point or the end point of the wall linestring; andD) in response to the modification to the wall linestring, automatically maintaining relative positions of the respective structural features along the wall segment by:preserving the percentage start value and the percentage end value for each wall span object of the one or more wall span objects, andcomputing updated geospatial coordinates for each respective structural feature by interpolating along a modified wall linestring according to the preserved percentage start value and the preserved percentage end value for each wall span object.

40. The method of claim 39, wherein each wall span object further includes one or more of a material designation indicating a material of the respective structural feature, a glass type indicating a type of glass for the respective structural feature when the span type is window,a roof height value indicating a height of a roof at the wall segment,one or more image references associating the respective structural feature with one or more images, andone or more observation references associating the respective structural feature with one or more inspection observations.

41. The method of claim 39, wherein the percentage start value and the percentage end value are each constrained to values in a range from 0 to 100, inclusive.

42. The method of claim 39, wherein the wall object is stored in walls data of a structure object, the structure object representing the building structure and being stored in structures data of a property features data structure, and wherein the property features data structure further includes trees data and an elevation section.

43. A system for managing property assessment data for wildfire risk calculation, the system comprising:a data store configured to store a property features data structure, the property features data structure including:Attorney Docket No. FWIG-003W001 structure data representing one or more building structures on a property, the structure data including, for each building structure of the one or more building structures, a structure shape defined as a polygon in geospatial coordinates, a structure type field indicating whether the building structure is a main structure or a fuel structure, wall data representing one or more wall segments of the building structure, and roof data representing a roof of the building structure;tree data representing one or more trees on or near the property, the tree data including, for each tree of the one or more trees, a tree shape defined as a polygon representing a canopy footprint, a height value, a canopy radius value, and trim information; andelevation data representing terrain elevation across the property and including geographic bounds and an plurality of elevation values;a landscape pipeline implemented by one or more first processors in communication with the data store, the landscape pipeline configured to:receive input data including an address of the property, satellite imagery of the property, and parcel data,generate an initial property features data structure using the input data, and store the initial property features data structure in the data store; a property application implemented by one or more second processors in communication with the data store, the property application configured to:display a rendering of the property based on the property features data structure,receive user input representing one or more corrections to the property features data structure, andgenerate one or more differential change records representing the one or more corrections; anda property ignition model implemented by one or more third processors in communication with the data store, the property ignition model configured to:receive the property features data structure as input, andcalculate a probability of destruction of at least one building structure of the one or more building structures due to wildfire based at least in part on the structure data, the tree data, and the elevation data.Attorney Docket No. FWIG-003W001 44. The system of claim 43, wherein the trim information for each tree of the one or more trees includes:a maximum radius reduction value representing a maximum amount by which the canopy radius value can be reduced by trimming,an actual trim height value representing a current trim height of the tree based on observed cut scars,a maximum trim height value representing a maximum height to which the tree can be trimmed, anda highest adjacent roofline value representing a height of a highest roofline of structures adjacent to the tree.

45. The system of claim 43, wherein each tree of the one or more trees further includes a breast height diameter value representing a diameter of a trunk of the tree measured at a standard forestry measurement height.

46. The system of claim 45, wherein the property ignition model is further configured to use the breast height diameter value to calculate an estimated heat production of the tree during combustion based on a correlation between trunk diameter and thermal energy output.

47. The system of claim 43, wherein the property ignition model is further configured to weight a tree of the one or more trees as a lower ignition risk when the tree is positioned at a higher elevation than the at least one building structure based on the elevation data.

48. The system of claim 43, wherein the tree data for each tree of the one or more trees further includes:a natural crown radius value representing an untrimmed canopy radius,a natural lowest branch height value representing a height of a lowest branch in an untrimmed state, anda species designation selected from a controlled vocabulary based on tree shape characteristics.

49. The system of claim 43, wherein the structure type field is set to one of:main, indicating the building structure is a primary structure to be protected from wildfire, orAttorney Docket No. FWIG-003W001 fuel, indicating the building structure is a neighboring structure or outbuilding that may contribute to wildfire ignition risk.

50. The system of claim 43, wherein the roof data for each building structure includes a roof object, the roof object including:a roof shape defined as a polygon in geospatial coordinates,materials data containing one or more roofing material classifications, and features data containing one or more roof feature objects.

51. The system of claim 50, wherein each roof feature object of the one or more roof feature objects includes a feature type field set to one of: chimney, skylight, or vent, and wherein when the feature type field is set to vent, the roof feature object further includes a vent type field and a vent subtype field.

52. The system of claim 43, wherein the wall data for each building structure includes one or more wall objects, each wall object including:a wall linestring defining a wall segment in geospatial coordinates, the wall linestring having a start point and an end point, andspans data containing one or more wall span objects, each wall span object representing a structural feature along the wall segment.

53. The system of claim 52, wherein each wall span object includes:a span type indicating a type of the structural feature,a percentage start value representing a starting position of the structural feature as a percentage of a length of the wall linestring, anda percentage end value representing an ending position of the structural feature as a percentage of the length of the wall linestring.

54. A property assessment apparatus for wildfire risk assessment, the apparatus comprising:one or more microprocessors;a display in communication with the one or more microprocessors; andnon-transitory computer memory in communication with the one or more microprocessors, the non-transitory computer memory storing:Attorney Docket No. FWIG-003W001 a property features data structure for a target property including a target building structure, the property features data structure including:an info section storing property metadata including an address of the target property and risk data aggregated from a plurality of authoritative sources,structures data storing one or more structure objects, each structure object including a structure shape defined as a polygon in geospatial coordinates, a structure type field, walls data, and a roof object, trees data storing one or more tree objects, each tree object including a tree shape defined as a polygon, a height value, a canopy radius value, a breast height diameter value, a species designation, and trim information,an elevation section storing a grid of elevation values with geographic bounds, andother shapes data storing one or more other shape objects representing miscellaneous property features; andcomputer-readable instructions that, when executed by the one or more microprocessors, cause the one or more microprocessors to:A) display, on the display, a geospatially registered rendering of the target property based on the property features data structure;B) receive user input representing a modification to the property features data structure;C) generate a differential change record representing the modification, the differential change record including a path specifying a hierarchical location within the property features data structure;D) apply the differential change record to the property features data structure to produce an updated property features data structure; andE) provide the updated property features data structure to a property ignition model for calculation of wildfire risk.

55. The apparatus of claim 54, wherein the computer-readable instructions, when executed by the one or more microprocessors, further cause the one or more microprocessors to:transform geometric data from WGS84 / EPSG:4326 geographic coordinates to EPSG:3857 projected coordinates for display on the display and for user interaction, andAttorney Docket No. FWIG-003W001 apply an inverse transformation to convert edited coordinates from EPSG:3857 projected coordinates back to WGS84 / EPSG:4326 geographic coordinates when storing changes to the property features data structure.

56. The apparatus of claim 54, wherein the property features data structure further includes an images section storing image metadata, the image metadata including:satellite imagery metadata including geographic bounds and resolution, and a reference to a three-dimensional model of the target building structure generated from wrap-around video using structure-from-motion techniques.

57. The apparatus of claim 56, wherein the three-dimensional model is stored in a Gaussian splatting format.

58. The apparatus of claim 54, wherein each other shape object of the other shapes data is classified as one of:a standalone feature having a position independent of the target building structure, or an attached feature having a position defined relative to the target building structure.

59. A system for aligning a three-dimensional reconstruction of a building structure with geospatial reference data, the system comprising:a data store configured to store:image data representing a three-dimensional reconstruction of the building structure, the three-dimensional reconstruction being generated from wrap-around video of the building structure using structure-from-motion techniques, and reference data including a geospatially calibrated polygon representing a footprint of the building structure;an alignment engine implemented by one or more processors in communication with the data store, the alignment engine configured to:A) determine an estimated perimeter of the building structure from the three- dimensional reconstruction;B) apply an iterative closest point technique to iteratively minimize a distance between the estimated perimeter and the geospatially calibrated polygon by applying successive rigid transformations, the successive rigid transformations including rotation, translation, and uniform scaling;Attorney Docket No. FWIG-003W001 C) determine that alignment is successful when the estimated perimeter, after application of the successive rigid transformations, falls within a threshold distance of the geospatially calibrated polygon; andD) store alignment data representing a final transformation resulting from the successive rigid transformations; anda virtual inspection engine in communication with the data store, the virtual inspection engine configured to display the three-dimensional reconstruction using the alignment data such that structural features identified in the three-dimensional reconstruction are geospatially registered to the geospatially calibrated polygon.

60. The system of claim 59, further comprising:a mobile device including an inertial measurement unit, the mobile device being configured to capture the wrap-around video of the building structure,wherein the alignment engine is further configured to receive time-synchronized inertial measurement unit data from the mobile device, the inertial measurement unit data including magnetometer readings for orientation, gyroscope data for angular velocity, and accelerometer data for linear acceleration.

61. The system of claim 60, wherein the alignment engine is further configured to use the inertial measurement unit data to estimate a camera pose and a camera trajectory during capture of the wrap-around video without requiring manually identified ground control points.

62. The system of claim 59, wherein the uniform scaling of the successive rigid transformations accounts for systematic scale errors in the three-dimensional reconstruction caused by focal length variations or scale drift in sequential feature tracking.

Images