Asset-level vulnerability and mitigation

Abstract

To obtain an insight into the hazard vulnerability of a parcel / asset.SOLUTION: There are provided methods, systems, and apparatus for receiving a request for a damage propensity score for a parcel, receiving imaging data for the parcel, wherein the imaging data comprises street-view imaging data of the parcel, extracting, by a machine-learned model including multiple classifiers, characteristics of vulnerability features for the parcel from the imaging data, determining, by the machine-learned model and from the characteristics of the vulnerability features, a damage propensity score for the parcel, and providing a representation of the damage propensity score for display.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] Cross - Reference to Related Applications This application claims priority to U.S. Patent Application No. 17 / 158,585, filed on January 26, 2021, entitled "ASSET - LEVEL VULNERABILITY AND MITIGATION", the disclosure of which is incorporated herein by reference.

Background Art

[0002] As land development encroaches on the boundary between wilderness and urban areas and environmental changes bring long - term droughts, wildfires are becoming an increasingly serious problem. Insurers and risk assessment managers look at the various assets present on a lot and use regression techniques and known vulnerabilities to generate wildfire risk assessments. Generating a risk assessment for a lot may require asset inspections and other one - time static appraisals, and as a result, changes to the lot can create a need for costly updated re - evaluations to continue to recognize the current risk.

Summary of the Invention

[0003] This specification describes systems, methods, devices, and other technologies related to using machine learning to gain insights about the hazard vulnerability of a lot / asset from imaging data that captures the lot / asset.

[0004] In general, one innovative aspect of the subject matter described in this specification can be embodied in a method of receiving a request for a damage propensity score for a lot and receiving imaging data for the lot, where the imaging data includes street - view imaging data of the lot. A machine - learning model that includes multiple classifiers extracts the characteristics of multiple vulnerability features of the lot from the imaging data and determines a damage propensity score for the lot from the characteristics of the multiple vulnerability features. A representation of the damage propensity score is provided for display.

[0005] These and other implementations may each optionally include one or more of the following features. In some embodiments, the method further includes generating a set of feature segments from the characteristics of a plurality of vulnerability features.

[0006] In some embodiments, this method further includes generating a three-dimensional model of a partition from the characteristics of multiple vulnerability features and imaging data of the partition.

[0007] In some embodiments, the imaging data of a section includes imaging data captured within a time threshold from the requested time.

[0008] In some embodiments, receiving a request for a damage tendency score includes receiving hazard event data for a hazard event and determining a partition damage tendency score for the hazard event from the characteristics of multiple vulnerability features and the hazard event data for the hazard event. The method may further include receiving updated hazard event data for the hazard event and determining an updated damage tendency score for the partition for the hazard event from the characteristics of multiple vulnerability features, the hazard event data, and the updated hazard event data.

[0009] In some embodiments, the method further includes determining one or more mitigation steps for a section using a machine learning model, determining an updated damage tendency score based on the one or more mitigation steps using a machine learning model, and providing a representation of the one or more mitigation steps and the updated damage tendency score. The one or more mitigation steps may include adjustments to the characteristics of multiple vulnerability features extracted from imaging data.

[0010] In some embodiments, determining one or more mitigation steps further includes iterating through updated damage tendency score determinations based on the tuned characteristics of multiple vulnerability features. In some embodiments, determining the updated damage tendency score further includes determining whether the updated damage tendency score satisfies a threshold damage tendency score.

[0011] In some embodiments, determining one or more mitigation steps includes determining one or more mitigation steps for a particular type of hazard event, wherein one or more mitigation steps for a first type of hazard event are different from one or more mitigation steps for a second type of hazard event.

[0012] In some embodiments, the method generates training data for a machine learning model, which further includes: generating training data for a machine learning model, which includes: receiving a plurality of parcels located in the vicinity of a hazard event, wherein each parcel of the plurality of parcels has received at least a threshold exposure to the hazard event; receiving parcel imaging data, including Street View imaging data, for each parcel of the plurality of parcels; and extracting from the imaging data the characteristics of a plurality of vulnerability features of a first subset of parcels of the plurality of parcels that did not burn during the hazard event and a second subset of parcels of the plurality of parcels that did burn.

[0013] In some embodiments, extracting the characteristics of multiple vulnerability features includes providing imaging data to multiple classifiers. Extracting the characteristics of multiple vulnerability features may include the identification of multiple objects in the imaging data by the multiple classifiers.

[0014] In some embodiments, the method further includes receiving additional structural characteristics for each section of a plurality of sections, extracting from the additional structural characteristics a second set of vulnerability features for a first subset of sections of the plurality of sections that did not burn during the hazard event and a second subset of sections of the plurality of sections that did burn, and providing the second set of vulnerability features to a machine learning model.

[0015] In some embodiments, additional structural features include post-hazard event inspection of multiple compartments.

[0016] The disclosure also provides a non-temporary computer-readable storage medium coupled to one or more processors that stores instructions, which, when executed by one or more processors, cause one or more processors to perform operations in an implementation of the methods provided herein.

[0017] It will be understood that the methods and systems provided herein may include any combination of the embodiments and features described herein. That is, the methods and systems provided herein are not limited to the combinations of embodiments and features specifically described herein, but may also include any combination of the embodiments and features provided herein.

[0018] Certain embodiments of the subject matter described herein can be implemented to achieve one or more of the following advantages. The advantage of the technology is that it can develop a novel understanding of hazard vulnerabilities for substantially more vulnerability features than conventional methods using trained machine learning models that consider the configuration of vulnerability feature properties depending on a specific set of hazard conditions and degrees of exposure, which may be more complex than the sum of risk factors and can reflect non-trivial features that contribute to the degree of damage suffered or damage / no damage outcome. Assessment of hazard vulnerabilities for specific hazards and degrees of exposure can be determined for a parcel using imaging data and may not require additional asset inspection. Hazard vulnerability assessments can be used when determining asset valuation, sales, taxes, etc.

[0019] By utilizing street view images of a parcel, it is possible to gain access to unique features of the parcel that are not otherwise available using other imaging data, such as features that reflect the current state of houses on the parcel (e.g., vines growing on the side of a house, the location of cars parked on the driveway). For example, vulnerability trend scores can be determined under real-time hazard conditions, where mitigation responses can be updated as hazard conditions change, in order to identify vulnerable parcels based on the respective vulnerabilities of each parcel under the current conditions of a hazard event. Optimized mitigation steps, such as risk reduction plans and / or cost-benefit estimates, can be determined in an iterative process by a trained machine learning model based on the extracted characteristics of the parcel's vulnerability features and in response to hazard events.

[0020] The applications of this technology generally include insurance risk assessment, real-time risk assessment and response, and general natural disaster hazard assessment and mitigation. More specifically, this technology can be used by local, state, or central governments to perform more accurate risk assessments and to design and develop risk mitigation plans.

[0021] Details of one or more embodiments of the subject matter described herein are shown in the accompanying drawings and in the following description. Other features, aspects, and advantages of the subject matter will become apparent from the description, drawings, and claims. [Brief explanation of the drawing]

[0022] [Figure 1] This is a block diagram of an exemplary operating environment for hazard vulnerability system 102. [Figure 2A] This shows satellite / aerial imagery including multiple exemplary sections before and after a hazard event. [Figure 2B] This shows street view-based images of an exemplary neighborhood before and after a hazard event. [Figure 2C] This shows Street View-based images of another exemplary parcel before and after a hazard event. [Figure 3] It is a flowchart of an exemplary process of a hazard vulnerability system. [Figure 4] It is a flowchart of another exemplary process of a hazard vulnerability system. [Figure 5] It is a block diagram of an exemplary computer system.

[0023] Similar reference numbers and names in various drawings indicate similar elements. **DETAILED DESCRIPTION**

[0024] Overview The technology of this patent application aims to use machine learning to obtain insights into the hazard vulnerability of a compartment / asset from imaging data that captures the compartment / asset.

[0025] More specifically, the technology of this application uses a trained machine learning model to identify vulnerability features and identify the characteristics of vulnerability features in the imaging data of a compartment, such as street view imaging data, LIDAR data, high-resolution satellite image data, aerial image data, infrared image data, user-provided images, etc. The characteristics of the vulnerability features can be used to generate a vulnerability tendency score and / or identify a mitigation strategy for reducing the hazard vulnerability of a specific compartment according to a specific hazard and the degree of exposure.

[0026] Generating training data for training the machine learning model can include selecting a set of compartments located within the vicinity of a hazard event, such as within the radius of a burn scar. The hazard event can be, for example, wildfire, flood, tornado, etc., and each set of compartments has experienced a certain degree of exposure to the hazard event. For each compartment in the set of compartments, imaging data that captures the compartment before the hazard event, such as a photo of a house / asset before experiencing a wildfire, is collected.

[0027] The vulnerability features of a parcel can be defined using existing risk assessment data, such as defendable space, building structure, or other features known to be associated with an increase or decrease in hazard vulnerability. Vulnerability features can be further extracted from imaging data showing parcels with damaged / undamaged and / or damaged outcomes for a particular hazard event, and one or more neural networks can be used to process the imaging data and extract additional vulnerability features determined to distinguish between damaged / undamaged and / or damaged outcomes for a parcel.

[0028] Vulnerability features, such as roof structure materials, distance between trees and houses, manufacturing information for building materials, frontage, fence type, and irrigation, can be extracted from parcel imaging data using multiple classifiers and object recognition techniques. Training data can be generated for multiple sets of parcels and their respective hazard events, and can include extracted vulnerability features, parcel location relative to hazard events, degree of exposure / damage during hazard events, and damaged / undamaged results. In addition, public records of parcels and information about hazard events can be used when generating training data for training machine learning models.

[0029] A trained machine learning model can receive requests for vulnerability assessments of specific parcels and requests for hazard events, including the degree of exposure. Imagery data can be collected about a parcel using, for example, its known address, geographical location, etc., and may include only imagery data captured within a time threshold, e.g., within the last six months. The imagery data can reflect the current state of the parcel, e.g., the current state of the surrounding vegetation, vehicle locations, structures built within the parcel (e.g., sheds), etc. The machine learning model can receive imagery data, public records (e.g., construction year, setbacks, special permits, etc.), and other relevant geospatial information (e.g., neighborhood housing density, distance to fire stations / emergency services, distance to major roads, etc.) as input, and extract the characteristics of the parcel's vulnerability features. The determined vulnerability trend score can be provided as output.

[0030] In some embodiments, the model can determine mitigation steps to reduce the risk score for a parcel based on the parcel's vulnerability features. Determining mitigation steps by a machine learning model may involve identifying the characteristics of vulnerability features that can be adjusted (e.g., trimming vegetation, changing roofing materials, changing siding materials) and iterating through risk score determination based on the adjusted characteristics of the vulnerability features. Permutations of mitigation steps can be evaluated for various hazard event scenarios to provide an optimized subset of mitigation steps for a particular parcel.

[0031] In some embodiments, real-time hazard vulnerability can be determined for a particular parcel based on real-time hazard events or potential future hazard risks, such as ongoing wildfires, droughts, or severe weather patterns. As a hazard event progresses, for example, as the degree of exposure changes, the parcel's hazard vulnerability is updated, and a real-time alert is generated accordingly to notify, for example, homeowners or emergency responders of the real-time hazard vulnerability and / or mitigation measures.

[0032] Exemplary operating environment Figure 1 is a block diagram of an exemplary operating environment 100 for the hazard vulnerability system 102. The hazard vulnerability system 102 can be hosted on one or more local servers, cloud-based services, or a combination thereof.

[0033] The hazard vulnerability system 102 can communicate with a network and data, and the network can be configured to enable the exchange of electronic communications between devices connected to the network. The network may include, for example, one or more of the following: the Internet, Wide Area Networks (WANs), Local Area Networks (LANs), analog or digital wired and wireless telephone networks (e.g., public switched telephone networks (PSTNs), Integrated Services Digital Networks (ISDNs), cellular networks, and Digital Subscriber Lines (DSLs)), wireless, television, cable, satellite, or any other distribution or tunneling mechanism for carrying data. The network may include multiple networks or subnetworks, each of which may include, for example, wired or wireless data paths. The network may include circuit-switched networks, packet-switched data networks, or any other networks capable of carrying electronic communications (e.g., data or voice communications). For example, a network may include a packet-switched network based on the Internet Protocol (IP), asynchronous transfer mode (ATM), PSTN, IP, X.25, or Frame Relay, or other equivalent technologies, and may support voice using, for example, VoIP or other equivalent protocols used for voice communication. A network may include one or more networks, including a wireless data channel and a wireless voice channel. A network may be a wireless network, a broadband network, or a combination of a wireless network and a broadband network.

[0034] The hazard vulnerability system 102 includes a training data generator 104 and a damage tendency model 106. Optionally, the hazard vulnerability system 102 includes a mitigation engine 108. The training data generator 104, the damage tendency model 106, and the mitigation engine 108 are described herein with reference, but the described operations can be performed by more or fewer subcomponents.

[0035] The training data generator 104 includes a vulnerability feature extractor 110, a parcel hazard event module 112, and a vector generator 114. The training data generator 104 receives imaging data 116 from a repository such as satellite and / or aerial imagery 118 and Street View imagery 117, and provides training data 120 as output. The output training data 120 can be used to train the damage trend model 106.

[0036] The damage trend model 106 includes multiple classifiers 107, for example, one or more neural networks or machine learning models, such as a random forest. The classifiers can be configured to classify the damage trend as a binary outcome, for example, damaged or undamaged, or they can be configured to classify the degree of the damage trend using, for example, a regression task. In some embodiments, the classifiers can be used to estimate damage to specific subcomponents of a plot, such as the roof of a building or the siding of a building, to further refine the damage trend model 106.

[0037] The damage trend model 106 can receive training data 120 containing a significant number of training vectors generated using a large sample of different hazard events documented in the imaging data 116 and historical hazard event data 124. The damage trend model 106 can be trained to infer damage trends for a particular parcel, partly based on the characteristics of the parcel's vulnerability features. Multiple classifiers 109, including the same or different sets of classifiers described, for example with reference to classifier 107, can process the received imaging data 116 to identify and label the characteristics of vulnerability features extracted from the imaging data 116.

[0038] The satellite / aerial imagery 118 includes any imagery that captures a geographical area and provides information about that geographical area. Information about the geographical area may include, for example, information about one or more parcels located within the geographical area, such as structures, vegetation, topography, etc. The satellite / aerial imagery may be, for example, Landsat imagery or other forms of aerial imagery. The satellite / aerial imagery 118 may be, for example, RGB imagery or hyperspectral imagery. The satellite / aerial imagery 118 can be captured using satellite technology, such as Landsat, or drone technology. In some implementations, the satellite / aerial imagery can be captured using other high-altitude technologies, such as drones, weather balloons, or airplanes. In some embodiments, synthetic aperture radar (SAR) imagery can be used in addition to satellite imagery as described herein.

[0039] In some implementations, satellite imagery or other aerial imagery can be captured using radar-based imaging, such as LiDAR imagery, RADAR imagery, or other types of imaging that use the electromagnetic spectrum, or a combination thereof. Satellite / aerial imagery 118 may include images of geographical areas containing various natural features, including different terrains, vegetation, bodies of water, and other features. Satellite / aerial imagery 118 may also include images of man-made developments, such as housing construction, roads, dams, and retaining walls.

[0040] The Street View images 117 include, for example, any images capturing aspects of one or more parcels from a frontage viewpoint, captured from a road or sidewalk facing the parcel. In some embodiments, the Street View images 117 can be captured by one or more cameras mounted on a vehicle and configured to capture Street View images 117 of the parcel as the vehicle passes through it. Optical and LiDAR Street View images 117 can be used to capture depth information about the parcel. In some embodiments, the Street View images 117 can have high spatial resolution; for example, the Street View images can have a spatial resolution of less than 1 centimeter.

[0041] In some embodiments, Street View images 117 can be captured by a user, for example, the homeowner of the parcel, from a frontage view of the property. Street View images can be captured using a built-in camera on a smart device, such as a smartphone or tablet, and / or using a handheld camera. In some embodiments, Street View images can be captured by a land surveyor, insurance appraiser, parcel appraiser, or other person documenting the characteristics of the parcel.

[0042] In some embodiments, the Street View image 117 may include additional views of the parcel, such as a side view of the parcel or a view captured from the rear of the parcel. For example, a user may capture a Street View image 117 of the backyard of their property, or a side view of their property.

[0043] In some embodiments, the imaging data 116 may include images of a section before and after a hazard event, for example, before and after a wildfire. The imaging data 116 associated with a particular hazard event may include burn marks, for example, areas damaged by the hazard event. A further explanation of the imaging data 116 is presented with reference to Figures 2A to 2C.

[0044] The training data generator 104 receives parcel data 122 from the repository of historical hazard event data 124. The parcel data 122 may include, for example, insurance assessments, land surveys, appraisals, building / construction records, code inspections / violations, and other public records.

[0045] In some embodiments, parcel data 122 may include public records from post-hazard event damage reports, such as from post-hazard insurance inspections, from a Damage Inspection (DINS) database maintained by CalFIRE. Post-hazard event damage reports may include direct inspections of structures exposed to a hazard event, such as a wildfire, and may include information on the vulnerability characteristics of the structure, such as roof type, eaves type, building materials, etc., as well as the level of damage. Parcel data 122 may include damage / non-damage data and / or degree of damage data for parcels within the radius of a hazard event, for example, within the radius of a burning radius.

[0046] The training data generator 104 receives image data 116 as input. The image data 116 captures one or more parcels at a specific location, for example, one or more houses, and at a specific time, for example, before or after a hazard event. For example, Street View images 117 can capture a house at a specific location address and a first date / time, for example, before a hazard event.

[0047] The vulnerability feature extractor 110 may include multiple classifiers 107, each capable of identifying vulnerability features, such as objects, within the imaging data 116. For each parcel shown in the imaging data 116, the vulnerability feature extractor 110 can extract vulnerability features and provide the parcel's vulnerability features F1, F2, ... FN as output to the vector generator module 114. Continuing the above example, the vulnerability features F1, F2, ... FN are extracted for a house at a specific location address and a first date / time, and may include, for example, roof structure, vegetation, frontage, and land slope.

[0048] Vulnerability features may include, but are not limited to, building materials, defendable spaces, plot gradients, and proximity to roads. Vulnerability features may further include objects, such as trees and vehicles. In some embodiments, ground truth labeling can be used to identify vulnerability features, for example, by human experts or in an automated / semi-automatic manner. Vulnerability features utilized by insurance adjusters / risk assessment managers can be identified in the imaging data 116.

[0049] Vulnerability features may further include vulnerability features of the parcel captured within the imaging data 116 that were not previously identified as risk hazards by insurance assessors / risk assessments. In other words, parcel features that may not have been previously labeled as hazard risks can be extracted as potential vulnerability features to generate training data. Examples include the structure of the driveway, the distance between a parked car and the house, and the type of grass seed used for the lawn. The damage tendency model 106 can define vulnerability features extracted from the imaging data 116 that may not otherwise be predicted to be significant by conventional means as significant, thereby allowing the vulnerability features to be processed by the machine learning model to determine which of the potential vulnerability features are significant in the parcel's damage tendency, for example, whether they have a statistical effect on the damaged / undamaged outcome and / or the degree of damage outcome. In this way, novel and non-trivial features can be identified as significant in the damage tendency.

[0050] Each vulnerability feature of a parcel describes the parcel's characteristics as shown in the imaging data 116, for example, in the Street View images 117 and / or satellite / aerial images 118. For each vulnerability feature extracted from the imaging data 116, one or more characteristics C1, C2, ... CN, e.g., F1{C1, C2, ... CN} of the vulnerability feature are extracted. Further details of feature extraction are described with reference to Figures 2A to 2C. The characteristics of a vulnerability feature may include quantifiable and / or qualitative aspects of the vulnerability feature. For example, a vulnerability feature that is a roof structure may be characterized by building materials, roof board spacing, age of construction, and roof maintenance. In another embodiment, a vulnerability feature that is a parcel slope may be characterized by a slope measurement of 0.5°.

[0051] The partition hazard event module 112 receives, as input, historical hazard event data 124 containing records of past hazard events and partition data 122 for each partition shown in the imaging data 116 processed by the vulnerability feature extractor 110. The partition hazard event module 112 provides the partition data 122 for each partition to the vector generator 114.

[0052] The historical hazard event data 124 may include the time of the hazard event, for example, the start time and end time of the hazard event. For example, the start time when a wildfire began, and the end time when the wildfire was completely contained or extinguished. The historical hazard event data 124 may also include the geographical location of the affected area, for example, the area including burn marks, for example, GPS coordinates.

[0053] The parcel data 122 for a specific parcel may include public records of the parcel before and after a hazard event, for example, before and after a wildfire. For example, parcel data 122 may include post-hazard insurance / assessment records, for example, damage assessments since the hazard event. In other words, parcel data 122 may include damage / non-damage results for a parcel in response to a specific hazard event, for example, damaged vs. undamaged. The parcel hazard event module 112 can use the damage / non-damage results and / or damage degree results of multiple parcels located within the burning radius 208 of a hazard event as ground truth for the training data 120.

[0054] In some embodiments, the house plot data 122 may include structural characteristics of the plot collected before the hazard event, such as building records, building materials, roof type, etc.

[0055] The training data generator 104 can generate training data from parcel images using imaging data 116 that occurs before and after a hazard event, and from parcel data 122 of parcels from historical hazard event data 124 corresponding to the event, for example, a wildfire. The vulnerability feature extractor 110 can extract vulnerability features and related characteristics of parcels in the imaging data 116 that each appear within the radius of a hazard event, for example, within the distance of a burn mark.

[0056] The vector generator 114 receives vulnerability features and characteristics of vulnerability features extracted from the feature extraction module 110, and optionally receives parcel data 122 from the parcel hazard event module 112 for a specific parcel as input. In some embodiments, the parcel data 122 can be used as ground truth for the training data 120, and parcels can be labeled as either "burned" or "unburned" using the parcel data 122, which includes damaged / undamaged results and / or damage degree results for hazard events.

[0057] The vector generator 114 can generate training data 120 from the extracted vulnerability features, the characteristics of the vulnerability features, and the partition data 122 for each partition, for example, each training vector V. Further details of the training data generation will be explained with reference to Figure 4.

[0058] The damage trend model 106 can receive the training data 120 as input to train a machine learning model, for example, the damage trend model 106. In some implementations, the damage trend model 106 can be trained using a considerable number of training vectors generated using large samples at different locations and partition data representing various historical hazard events. In one embodiment, the training data 120 provided to the damage trend model 106 may include thousands of partitions that experience many different hazard events.

[0059] The hazard vulnerability system 102 receives a request 126 from a user on a user device 128. The user device 128 can include, for example, a mobile phone, a tablet, a computer, or another device including an operating system 129 and an application environment 130 through which the user can interact with the hazard vulnerability system 102. In one embodiment, the user device 128 is a mobile phone including an application environment 130 configured to display a view 132 including at least a portion of a parcel. In one embodiment, as shown in Figure 1, the application environment 130 displays a view 132 including a street view of a house and surrounding assets such as trees, bushes, shrubs, etc.

[0060] Request 126 may include the location of a parcel specified by the user of the user device 128. The parcel location may include a geographical location, such as GPS coordinates or a local address, and can be entered by the user into the application environment 130.

[0061] Request 126 may further include a request for a damage tendency score, for example, a request for relative vulnerability to a hazard event, and Request 126 may specify a specific hazard event, for example, a specific real-time hazard event, or a general type of hazard event, for example, a wildfire, flood, earthquake, etc. In one embodiment, a user may submit Request 126 specifying the location address of a parcel and request a damage tendency score for that parcel to a real-time hazard event, for example, an ongoing wildfire. In another embodiment, a user may submit Request 126 specifying the location of a parcel, including GPS coordinates, and request a flood-specific damage tendency score for the parcel.

[0062] In some embodiments, the damage tendency score may be a relative measure of the risk that a particular parcel will be damaged by a hazard event. The damage tendency score may be a general measure of the risk to a particular parcel being damaged by a certain type of hazard event, such as a wildfire, or it may be a specific measure of the risk to a particular parcel being damaged by a specific hazard event, such as a real-time flood event.

[0063] In some embodiments, the damage tendency score may include a percentage of loss for a given parcel under a specific hazard scenario, e.g., a percentage of damage. For example, the damage tendency score could be 10% loss for a particular parcel under a specific wildfire scenario.

[0064] In some embodiments, end users, such as asset owners, insurance assessors, and government officials, can provide a complete stochastic hazard model, i.e., a distribution of hazard characteristics and associated probabilities, thereby enabling the hazard vulnerability system to provide the expected average annual loss (AAL) for a particular parcel.

[0065] In some embodiments, requirement 126 can specify a location comprising multiple sections, such as a neighborhood, a street comprising multiple houses, or a complex comprising multiple buildings. The user may be interested in determining an individual damage tendency score for each structure in a location comprising multiple sections, or in determining a global damage tendency score for multiple sections.

[0066] In some embodiments, the hazard vulnerability system 102 receives as input a request 126 which includes a request for mitigation steps to reduce the hazard vulnerability of a plot. The mitigation engine 108 receives the request for mitigation steps and, based on imaging data 116 and damage tendency score 140, can identify a set of mitigation steps 136 that a user can take to reduce the hazard vulnerability of a plot. Mitigation steps are quantifiable and / or quantifiable measures that a user, e.g., a homeowner, can take to reduce the damage tendency score 140. Mitigation steps may include, for example, the removal / reduction of vegetation adjacent to a structure, building materials used on a structure, etc. For example, a mitigation step may be trimming leaves within a 2-foot radius surrounding a house. In another embodiment, a mitigation step may be changing the siding material of a house. In yet another embodiment, a mitigation step may be digging an irrigation ditch to collect runoff from a flooded stream area.

[0067] In some embodiments, the mitigation step can be provided within an application environment 130 on a user device 128, and the mitigation step 136 is visually identified, for example, by an indicator being overlaid on a view 132 of the area. For example, a tree with branches overhanging a roof can be visually identified within the application environment, for example, by using a box surrounding the tree and / or branches.

[0068] In some embodiments, the mitigation engine 108 can receive real-time event data 142, for example, real-time data about an occurring hazard event, and update the mitigation step 136 in real time to provide the user with a real-time response to the hazard event. The real-time event data 142 may include, for example, the extent of the hazard, weather patterns, mitigation events, and emergency responses. For example, real-time event data 142 for a wildfire may include the real-time perimeter of the fire, the percentage of firefighters controlling it, evacuation data, and wind warnings. In another embodiment, real-time event data 142 for a flood may include real-time river / stream levels, flood levels, rain / weather forecasts, and evacuation data.

[0069] Feature extraction As described above with reference to Figure 1, the vulnerability feature extractor 110 can receive imaging data 116, including satellite / aerial imagery 118 and Street View imagery 117, as extracted vulnerability features having input and associated properties. Figure 2A is a schematic diagram of an exemplary pair of satellite images containing multiple parcels before and after a hazard event. Satellite images 200a and 200b are captured at capture times T1 and T2, respectively, where T1 is the time occurring before the hazard event and T2 is the time occurring after the hazard event. In addition, T1 and T2 can be selected based on parcel data 122 of the parcels contained in satellite images 200a and 200b, where the parcel data includes records of the parcels at time T1' before the hazard event and time T2' after the hazard event.

[0070] Satellite images 200a and 200b show the same geographical area 202, including a set of parcels 204. Satellite image 200a is captured at time T1, which falls within a first time threshold prior to the onset of the hazard event, and satellite image 200b is captured at time T2, which falls within a second time threshold after the end of the hazard event.

[0071] Satellite imagery 200b captured after a hazard event includes a burn mark 206 resulting from the hazard event, such as a fire. The burn mark 206 may indicate an area of ​​geographic region 202 damaged / affected by the hazard event. The burn radius 208 defines a perimeter that includes an additional area encompassing, surrounding, and buffering the burn mark 206. The burn radius 208 may include additional radii extending outward from the burn mark, such as an additional 100 feet, an additional 1000 feet, and an additional 5000 feet. The burn radius 208 may include a section A damaged / affected by the hazard event and a section B not damaged / affected by the hazard event. Section A may be a section located within the burn mark 206 and damaged / affected by the hazard event. Section B may be a section located outside the burn mark 206 but within the burn radius 208, or Section B may be a section located within the burn mark 206 but not damaged / affected by the hazard event.

[0072] Satellite imagery 200b captured after a hazard event may include multiple burn marks 206 and burn radii, and the burn marks and / or burn radii may overlap with each other. Sections may be located in overlapping areas of burn marks and / or burn radii.

[0073] As illustrated with reference to Figure 1, the vulnerability feature extractor 110 receives satellite images 200a and 200b and extracts vulnerability features F1, F2, ..., FN and the characteristics of each vulnerability feature from the images 200a and 200b. Vulnerability features extracted from satellite images 200a and 200b may include, for example, roof structures, the location of a parcel relative to natural formations (e.g., forest / tree coverage, waterways, etc.), and the location of a parcel relative to artificial features (e.g., roads, irrigation ditches, farmland, etc.). Each characteristic may include, for example, building materials for roof structures, such as ceramics, metals, wood, etc. In another embodiment, the characteristic of the parcel's location relative to artificial features may include the relative distance between the parcel and the artificial features, for example, the distance from the house to the street.

[0074] In some embodiments, vulnerability features can be extracted from satellite imagery 118 for multiple parcels appearing in the satellite imagery. Additional vulnerability features can be extracted for each parcel of the multiple parcels using higher-resolution images, for example, using Street View images 117.

[0075] Figure 2B is a schematic diagram of an exemplary pair of Street View images of parcel A captured before and after a hazard event. Street View images 220a and 220b show a street view of the parcel, captured from street level, of, for example, a house 222 and surrounding vegetation 224 located on parcel A. Street View image 220b, captured after the hazard event, includes, for example, hazard event damage 226 to the house, an adjacent storage shed, and a nearby tree.

[0076] Vulnerability features are extracted from Street View images 220a and 220b. As illustrated with reference to Figure 1, the vulnerability feature extractor 110 receives imaging data 116, including Street View images 117, e.g., images 220a and 220b, and extracts vulnerability features F1, F2, ..., FN. Referring back to Figure 2B, the vulnerability features extracted from Street View images 220a and 220b include a cluster of trees (F1), a cluster of bushes (F2), an outdoor shed (F3), and the roof structure of house 222 (F4). More or fewer vulnerability features can be extracted from Street View images 220a and 220b, and the examples provided are not limiting.

[0077] For each vulnerability feature extracted from Street View images 220a and 220b, the vulnerability feature extractor identifies the characteristics of each vulnerability feature. For example, a group of trees F1 may have associated quantifiable characteristics such as the distance of the trees to house 222, the number of trees, the height of the trees, the proximity of the group of trees, and other quantifiable characteristics such as the health of the trees and whether they are covered with ivy. In another embodiment, a roof structure F4 may have associated characteristics such as the building material, the eaves structure, the roof slope, the age of the roof, the maintenance of the roof, and the degree of gutter filling.

[0078] In some embodiments, the vulnerability feature extractor 110 can identify vulnerability features in section A, for example, in the Street View image 220b, that reflect damage from a hazard event. In other words, the vulnerability feature extractor 110 can focus on vulnerability features whose properties reflect damage as a result of a hazard event, such as roof structures showing burn / smoke damage, burnt wood, etc.

[0079] Figure 2C is a schematic diagram of an exemplary pair of Street View images of parcel B captured before and after a hazard event. Street View images 240a and 240b show street-level and parcel-facing Street Views of, for example, a house 242 and surrounding vegetation 224 located on parcel B, shown outside the burn marks 206 of the hazard event and within the burn radius 208. Unlike parcel A, which was described with reference to Figure 2B, parcel B in Figure 2C is shown as undamaged from the hazard event.

[0080] Similarly, as illustrated with reference to Figure 2B, vulnerability features extracted from Street View images 240a and 240b include a cluster of trees (F5), a cluster of bushes (F6), a frontage space between a house and the street (F7), and the roof structure of house 246 (F8). More or fewer vulnerability features can be extracted from Street View images 240a and 240b, and the examples provided are not limiting.

[0081] For each vulnerability feature extracted from Street View images 240a and 240b, the vulnerability feature extractor identifies the characteristics of each vulnerability feature. For example, a group of trees F5 may have associated quantifiable characteristics such as the distance of the trees to house 242, the number of trees, the height of the trees, the proximity of the group of trees, and other quantifiable characteristics such as the health of the trees and whether they are covered with ivy. In another embodiment, the frontage space F7 between a house and a street may have characteristics such as distance, slope, and type of ground cover (e.g., cement vs. grass).

[0082] The vulnerability feature extractor 110 provides the vector generator 114 with vulnerability features and characteristics extracted from the imaging data 216, for example, 200a, 200b, 220a, 220b, 240a, and 240b, in order to generate the training data 120.

[0083] Exemplary process Figure 3 is a flowchart illustrating an exemplary process of the hazard vulnerability system 102. System 102 receives a request for a parcel damage tendency score (302). The request 126 can be provided to the hazard vulnerability system 102 by a user through the graphical user interface of the application environment 130. The request 126 may include a request for a damage tendency score 140 for a particular parcel, and may further include a request for one or more mitigation steps 136 to reduce the damage tendency score 140. The request 126 may further specify a particular hazard event, such as an ongoing wildfire, or a common hazard event type, such as a flood, and request a damage tendency score accordingly.

[0084] The system receives parcel imaging data, including street view imaging data of the parcel (304). The hazard vulnerability system 102 can receive imaging data 116, including street view images 117, from the image data repository. Each image may include a specific designated parcel of interest to the user. In some embodiments, the user can capture additional street view images of the parcel with a camera, for example, the built-in camera of the user device 128, and upload them to the hazard vulnerability system 102.

[0085] In some embodiments, the system receives imaging data of captured sections within a threshold time from the time of request 126, for example, within 6 months, within 2 weeks, within 1 hour, etc.

[0086] A machine learning model including a classifier extracts the characteristics of the parcel's vulnerability features from the imaging data (306). The damage trend model 106 receives the imaging data 116 and can extract the characteristics of the vulnerability features using multiple classifiers 107. The vulnerability feature extractor 110 receives images of the parcel from the imaging data 116 and extracts the parcel's vulnerability features F1, F2, ... FN, and for each vulnerability feature FN, the vulnerability feature extractor 110 extracts one or more characteristics C1, C2, ... CN of the vulnerability feature FN. In one embodiment, the vulnerability feature extractor 110 receives Street View images of the parcel capturing the house and surrounding assets and uses multiple classifiers to identify vulnerability features, for example, the house, vegetation, frontage area, fence, etc. The vulnerability feature extractor 110 identifies each characteristic of the extracted vulnerability features, for example, the building materials used on the house, for example, siding type, roof type, eaves structure, etc.

[0087] The machine learning model determines a parcel propensity score from the characteristics of the vulnerability features (308). The damage propensity model 106 is trained on training data 120, which includes multiple hazard events, e.g., hundreds of hazard events, and multiple parcels for each hazard event, e.g., thousands of parcels, as will be further described below with reference to Figure 4, so that the model 106 can infer between the characteristics of the parcel vulnerability features and the damage propensity of the parcel. The model 106 generates a damage propensity score 140 as output.

[0088] The system provides a representation of the propensity score for display (310). The damage propensity score 140 can be provided within the graphical user interface of the application environment 130. The propensity score 140 can be visually represented, for example, as a numerical value, and / or presented with contextual cues, including, for example, a relative hazard scale and color coding (high, medium, or low risk). The propensity score 140 can be presented with contextual information to help the user better understand the significance of the propensity score 140.

[0089] In some embodiments, the hazard vulnerability system 102 can utilize the characteristics of the partition's vulnerability features and imaging data to generate a three-dimensional model of the partition. The three-dimensional model of the partition can be displayed in a graphical user interface to help the user understand the damage tendency score and / or one or more mitigation steps 136 for optimizing the damage tendency score.

[0090] In some embodiments, the hazard vulnerability system 102 may determine one or more mitigation steps 136 to provide to the user as a way to reduce risk, for example, by optimizing the damage tendency score 140. The mitigation engine 108 receives the damage tendency score and the characteristics of the vulnerability feature of the plot and may determine one or more mitigation steps 136. The mitigation steps 136 can be identified by the damage tendency model 106 based on inferences about which characteristics of the vulnerability feature reduce the damage tendency. Mitigation steps 136 may include, for example, trimming tree branches away from the house. In another embodiment, mitigation steps 136 may include changing the roofing material, changing the location of an outdoor shed, removing brushes from the frontage area of ​​the house, etc. The mitigation engine 108 may provide the proposed updated characteristics of the vulnerability feature to the damage tendency model 106. The damage tendency model 106 receives the proposed updated characteristics of the vulnerability feature and may determine an updated damage tendency score 140.

[0091] In some embodiments, the mitigation engine 108 can determine the mitigation step 136 by calculating the gradient of the damage tendency score 140 for each vulnerability feature vector. Vulnerability feature vectors having the threshold magnitude gradient and / or the subset of the largest magnitude gradients from the set of gradients can be used as the basis for the mitigation step 136. In other words, vulnerability features that have the greatest impact (larger magnitude gradient) on the damaged / undamaged and / or damaged outcome can be the focus of the mitigation step because they can have a greater impact on the damage tendency score 140 than vulnerability features that have a smaller impact (smaller magnitude gradient) on the outcome. For example, roofing structural materials can be selected as the mitigation step if the gradient of the damage tendency score 140 for the vulnerability feature vector for roofing structural materials is at least the threshold magnitude, or ranks highest among the gradients of the vulnerability feature vectors for the section.

[0092] In some embodiments, a neural network can be used to infer vulnerability features in the imaging data that have the greatest impact on the damage tendency score 140 determined by the damage tendency model 106.

[0093] In some embodiments, the graphical user interface of the application environment 130 may present a visual representation of the mitigation step 136 on the user device 128. For example, a visual indicator 138 may identify a mitigation step and indicate, for example, to remove vegetation from a plot. The visual indicator 138 may be a bounding box surrounding the identified mitigation step 136, a graphical arrow, or other indicator. The visual indicator 138 may include text-based information about the mitigation step 136, for example, explaining how and why the mitigation step 136 reduces the plot's damage tendency score 140.

[0094] In some embodiments, the system 102 can perform the optimization process by iterating over the proposed properties for the vulnerability features of the partitions and calculating an updated propensity score 140 until an optimized, for example, lowest damage propensity score is found. The optimization process can continue until a threshold damage propensity score is met.

[0095] In some embodiments, the hazard vulnerability system 102 can further incorporate a cost analysis of the mitigation step 136. In other words, the hazard vulnerability system 102 can determine a mitigation step 136 that balances the optimization of the damage tendency score 140 while keeping the cost of the mitigation step 136 below a threshold cost.

[0096] In some embodiments, the hazard vulnerability system 102 may determine one or more mitigation steps based on a specific type of hazard event, such as flood versus wildfire, where the mitigation steps for a first type of hazard event differ from those for a second type of hazard event. For example, mitigation steps corresponding to a flood hazard may include digging drainage ditches and updating the gutter system, while mitigation steps corresponding to a wildfire may include mowing the vegetation surrounding a house.

[0097] In some embodiments, the hazard vulnerability system 102 receives a request 126 for a real-time damage propensity score 140 corresponding to a real-time hazard event. The hazard vulnerability system 102 receives real-time event data 142 about the hazard event and can determine the propensity score of the parcel corresponding to the hazard event from the characteristics of the vulnerability feature and the real-time event data. The hazard vulnerability system 102 can re-evaluate the propensity score 140 in real time based in part on updated hazard event data about the hazard event, such as changes in weather conditions or containment, and provide the user with the updated propensity score 140 according to the updated hazard event data.

[0098] In some embodiments, the mitigation engine 108 receives real-time event data 142. This real-time event data can be used to provide a user with real-time mitigation steps 136 via a user device 128 to reduce a parcel damage score 140 corresponding to an ongoing hazard event. For example, real-time event data including wildfire spread and containment, weather patterns, and emergency responder alerts can be used to help homeowners take immediate action to address wildfire spread and reduce the likelihood of their property being damaged by wildfire.

[0099] Figure 4 is a flowchart of another exemplary process of the hazard vulnerability system. The hazard vulnerability system 102 generates training data for a machine learning model (402). Training a machine learning model, for example, a damage tendency model 106, involves generating training data 120, which includes a large sample set of imaging data 116 and partition data 122, for example, historical hazard event data 124 containing thousands of images for hundreds of hazard events. The generated training data represents various imaging conditions, such as weather conditions, lighting conditions, seasons, etc., for various hazard events, such as hazards of different scales, hazard extent, location, and hazard type, and the damage tendency model 106 trained on the training data 120 can be generalized to develop heuristics for a wide range of imaging conditions and hazard events.

[0100] The system receives, for a hazard event, a plurality of compartments located within the vicinity of the hazard event, where each compartment has received at least a threshold exposure to the hazard event (404). The system can receive historical hazard event data 124 for the hazard event, which may include compartment data 122 for each of the plurality of compartments located within the vicinity of the hazard event, for example, within the combustion radius 208.

[0101] Proximity to a hazard event can be defined as being located within a combustion radius 208 surrounding the burn mark 206. As shown in Figure 2A, the combustion radius 208 can define an extended area surrounding the burn mark 206. In some embodiments, proximity to a hazard event may be a threshold distance from the outer perimeter of the burn mark 206, for example, within 1 mile, within 100 feet, within 5 miles, etc.

[0102] Threshold exposure is the minimum amount of exposure a compartment receives to a hazard event, and can be defined, for example, by the distance of the compartment to the hazard event, by the time the compartment is actively exposed to the hazard event, such as the time the asset is actively exposed to a wildfire. In some embodiments, threshold exposure can be defined using an emergency responder metric, such as a high-risk or evacuation zone considered by the emergency responder. In one embodiment, a compartment can satisfy threshold exposure by being considered to be within a wildfire evacuation zone. In another embodiment, a compartment can satisfy threshold exposure by having at least a portion of it come into contact with floodwaters (or a wildfire, or a tornado, or an earthquake, etc.). In yet another embodiment, a compartment can satisfy threshold exposure by being located within or near a threshold of a burn mark.

[0103] In some embodiments, each of the multiple compartments located within the combustion trace can be considered to have received a threshold exposure.

[0104] In some embodiments, the fire radiative power (FRP) of a fire can be calculated from mapped remotely perceived derived measurements of fire intensity. For example, using satellite data of an active fire, any structure located within a specific area and having a given threshold FRP value can be counted as experiencing the same threshold exposure.

[0105] The system receives imaging data for each of a plurality of parcels located within the vicinity of a hazard event, including Street View imaging data (406). The imaging data 116 may include satellite / aerial images 118 and Street View images 117 collected by the hazard vulnerability system 102 from a repository of collected images located in various databases and sources. Each image in the imaging data 116 includes the capture time when the image was captured and the geographical area containing that parcel among the plurality of parcels. The satellite / aerial image 118 includes the geographical area captured at a specific resolution and location information, such as GPS coordinates, that defines the geographical area captured within the image frame. The Street View image 117 includes street-level views of one or more parcels among a plurality of parcels captured at a specific resolution, such as one parcel, two parcels, etc., and location information, such as the address, that defines the location of the parcel captured in the Street View image 117.

[0106] The system extracts from the imaging data the characteristics of multiple vulnerability features of a first subset of sections that did not burn during the hazard event and a second subset of sections that did burn (408). As described with reference to Figures 1, 2A to 2C, the vulnerability feature extractor 110 can receive the imaging data 116 and extract vulnerability features using multiple classifiers. In some embodiments, extracting vulnerability features using multiple classifiers includes identifying objects in the imaging data 116 using multiple classifiers.

[0107] In some embodiments, the system can receive parcel data 122 of a first subset of parcels that did not burn during a hazard event and a second subset of parcels that did burn. The parcel data 122 may include additional structural characteristics for each parcel, such as post-hazard event inspections, building / construction records, appraisals, insurance assessments, etc. From the additional structural data, the system can extract vulnerability features and vulnerability feature characteristics for the first subset of parcels that did not burn and the second subset of parcels that burned.

[0108] The system can generate training vectors from the extracted vulnerability features and their characteristics. In some embodiments, the vector generator module 114 generates training vectors from the extracted vulnerability features and their characteristics for a machine learning model.

[0109] System 102 can generate training data 120 for a specific hazard event using the extracted vulnerability features and characteristics of each of the multiple sections. Damage / non-damage records and / or damage degree records for a specific section among the multiple sections from historical hazard event data 124 can be used as ground truth for the burn / non-burn outcome of each section.

[0110] The system provides training data to a machine learning model (410). The training data 120 is provided to a machine learning model, for example, a damage tendency model 106, to train the damage tendency model 106 to infer about the damage tendency of a particular section to a general hazard event type or a specific hazard event.

[0111] Figure 5 is a block diagram of an exemplary computer system 500 that can be used to perform the operations described above. The system 500 comprises a processor 510, memory 520, a storage device 530, and an input / output device 540. Each of the components 510, 520, 530, and 540 can be interconnected, for example, using a system bus 550. The processor 510 can process instructions to be executed within the system 500. In one implementation, the processor 510 is a single-threaded processor. In another implementation, the processor 510 is a multi-threaded processor. The processor 510 can process instructions stored in memory 520 or the storage device 530.

[0112] Memory 520 stores information within the system 500. In one implementation, memory 520 is a computer-readable medium. In another implementation, memory 520 is a volatile memory unit. In yet another implementation, memory 520 is a non-volatile memory unit.

[0113] The storage device 530 can provide mass storage to the system 500. In one implementation, the storage device 530 is a computer-readable medium. In various different implementations, the storage device 530 may include, for example, a hard disk device, an optical disk device, a storage device shared over a network by multiple computing devices (e.g., a cloud storage device), or some other mass storage device.

[0114] The input / output device 540 provides input / output operation to the system 500. In one implementation, the input / output device 540 may include one or more network interface devices, such as an Ethernet card, a serial communication device, such as an RS-232 port, and / or a wireless interface device, such as a 502.11 card. In another implementation, the input / output device may include a driver device configured to receive input data and transmit output data to other input / output devices, such as a keyboard, printer, and display device 560. However, other implementations, such as mobile computing devices, mobile communication devices, and set-top box television client devices, may also be used.

[0115] An exemplary processing system is shown in Figure 5, but the subject matter and functional operation implementations described herein can be implemented in other types of digital electronic circuits, or in computer software, firmware, or hardware, or a combination of one or more thereof, including the structures disclosed herein and their structural equivalents.

[0116] In this specification, the term “configured” is used in relation to systems and computer program components. One or more computer systems are configured to perform a particular operation or action to mean that the system has installed software, firmware, hardware, or a combination thereof that causes the system to perform the operation or action while in operation. One or more computer programs are configured to perform a particular operation or action to mean that one or more programs contain instructions that, when executed by a data processing device, cause the device to perform the operation or action.

[0117] The subject matter and functional operating embodiments described herein can be implemented in digital electronic circuits, tangibly embodied computer software or firmware, computer hardware, or one or more combinations thereof, including the structures disclosed herein and their structural equivalents. Embodiments of the subject matter described herein can be implemented as one or more modules of computer programs, i.e., computer program instructions encoded on a tangible non-temporary storage medium for execution by a data processing device or for controlling the operation of a data processing device. The computer storage medium may be a machine-readable storage device, a machine-readable storage board, a random or serial access memory device, or one or more combinations thereof. Alternatively or additionally, the program instructions may be encoded on artificially generated propagating signals, such as machine-generated electrical, optical, or electromagnetic signals generated to encode information to be transmitted to a receiving device suitable for execution by a data processing device.

[0118] The term "data processing device" refers to data processing hardware and encompasses all kinds of devices, machines, and equipment for processing data, including, for example, programmable processors, computers, or multiple processors or computers. A device may also be, or further include, dedicated logic circuits such as FPGAs (Field-Programmable Gate Arrays) or ASICs (Application-Specific Integrated Circuits). Optionally, in addition to hardware, a device may include code that creates an execution environment for computer programs, such as processor firmware, protocol stacks, database management systems, operating systems, or code comprising one or more of these.

[0119] A computer program, which may also be called or described as a program, software, software application, app, module, software module, script, or code, may be written in any form of programming language, including compiled or interpreted languages, or declarative or procedural languages, and may be deployed in any form, including standalone programs or modules, components, subroutines, or other units suitable for use in a computing environment. A program may, but is not required, correspond to a file in a file system. A program may be stored in a part of a file that holds other programs or data, for example, in a markup language document, in a single file dedicated to the program in question, or in a set of collaborative files, for example, in one or more scripts stored in a file that stores one or more modules, subprograms, or parts of code. A computer program may be deployed to run on one computer, or on multiple computers located in one site, or distributed across multiple sites and interconnected by a data communication network.

[0120] In this specification, the term “engine” is used broadly to refer to a software-based system, subsystem, or process programmed to perform one or more specific functions. Generally, an engine is implemented as one or more software modules or components and installed on one or more computers in one or more locations. In some cases, one or more computers are dedicated to a particular engine. In other cases, multiple engines may be installed and run on the same computer (one or more).

[0121] The processes and logic flows described herein can be executed by one or more programmable computers running one or more computer programs to perform their functions by acting on input data and producing outputs. The processes and logic flows can also be executed, for example, by dedicated logic circuits such as FPGAs or ASICs, or by a combination of application-specific logic circuits and one or more programmed computers.

[0122] A computer suitable for running computer programs may be based on a general-purpose or dedicated microprocessor, or both, or any other type of central processing unit. Generally, the central processing unit receives instructions and data from read-only memory, random-access memory, or both. Essential elements of a computer are a central processing unit for executing or running instructions, and one or more memory devices for storing instructions and data. The central processing unit and memory may be complemented by or incorporated into special-purpose logic circuits. Generally, a computer may also include one or more mass storage devices for storing data, such as magnetic disks, magneto-optical disks, or optical disks, or be operablely coupled to receive data from them, transfer data to them, or both. However, a computer is not required to have such devices. Furthermore, computers can be integrated into other devices, such as mobile phones, personal digital assistants (PDAs), mobile audio or video players, game consoles, Global Positioning System (GPS) receivers, or portable storage devices, such as Universal Serial Bus (USB) flash drives.

[0123] Computer-readable media suitable for storing computer program instructions and data include, for example, all forms of non-volatile memory, media, and memory devices, including semiconductor memory devices such as EPROM, EEPROM, and flash memory devices, magnetic disks such as internal hard disks or removable disks, magneto-optical disks, and CD-ROM and DVD-ROM disks.

[0124] To provide user interaction, embodiments of the subject matter described herein may be implemented on a computer, which may have a display device for displaying information to the user, such as a CRT (cathode ray tube) or LCD (liquid crystal display) monitor, and a keyboard and pointing device, such as a mouse or trackball, to which the user can provide input to the computer. Other types of devices may be used similarly to provide user interaction, for example, the feedback provided to the user may be any form of sensory feedback, such as visual feedback, auditory feedback, or tactile feedback, and input from the user may be received in any form, including acoustic, voice, or tactile input. In addition, the computer may interact with the user by sending documents to and receiving documents from devices used by the user, for example, by sending a web page to a web browser on the user's device in response to a request received from a web browser. The computer may also interact with the user by sending text messages or other forms of messages to a personal device, such as a smartphone running a messaging application, and receiving response messages from the user as replies.

[0125] Data processing equipment for implementing machine learning models may also include, for example, dedicated hardware accelerator units for handling the general, computationally intensive parts of machine learning training or production, i.e., inference workloads.

[0126] Machine learning models can be implemented and deployed using machine learning frameworks such as the TensorFlow framework, the Microsoft Cognitive Toolkit framework, the Apache Singa framework, or the Apache MXNet framework.

[0127] Embodiments of the subject matter described herein can be implemented in a computing system that includes, for example, a backend component as a data server, or a middleware component, such as an application server, or a frontend component, such as a graphical user interface, a web browser, or a client computer having an application that allows a user to interact with the implementation of the subject matter described herein, or any combination of one or more such backend, middleware, or frontend components. The components of the system can be interconnected by any form or medium of digital data communication, such as a communication network. Examples of communication networks include local area networks ("LANs") and wide area networks ("WANs"), such as the Internet.

[0128] A computing system can include clients and servers. Clients and servers are generally geographically separated and typically interact via a communication network. The client-server relationship arises from computer programs running on each computer that have a client-server relationship with each other. In some embodiments, the server sends data, such as an HTML page, to a user device for the purpose of displaying data to a user interacting with a device acting as a client and receiving user input from the user. Data generated on the user device, such as the results of user interactions, can be received from the device to the server.

[0129] This specification contains details of many individual implementations, which should be interpreted as descriptions of features specific to a particular embodiment, rather than limiting the scope of any feature or the scope of the claims. Certain features described herein in the context of a separate embodiment may also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may be implemented separately in multiple embodiments or in any preferred partial combination. Furthermore, features may be described above as acting in a particular combination, and may even be initially claimed as such, but one or more features from a claimed combination may, in some cases, be removed from the combination, and the claimed combination may be directed towards a partial combination or a variation of a partial combination.

[0130] Similarly, although the operations are depicted in a specific order in the drawings, this should not be understood as requiring that such operations be performed in a specific or sequential order shown, or that all exemplified operations be performed, in order to achieve the desired result. In certain circumstances, multitasking and parallel processing may be advantageous. Furthermore, the separation of various system components in the embodiments described above should not be understood as requiring such separation in all embodiments, and it should be understood that the described program components and systems can generally be integrated together in a single software product or packaged in multiple software products.

[0131] Accordingly, specific embodiments of the subject matter have been described. Other embodiments are within the scope of the following claims. In some examples, the operations described in the claims can be performed in a different order and still achieve the desired results. Furthermore, the processes shown in the accompanying drawings do not necessarily require the specific order or sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing may be advantageous.

Claims

1. Receiving a request for a section damage tendency score for one or more hazard event scenarios, Receiving imaging data of the said section, wherein the imaging data captures the characteristics of the said section, The process involves extracting the characteristics of multiple vulnerability features of the section from the imaging data using a trained machine learning model that includes multiple classifiers, The trained machine learning model determines the damage tendency score of the partition for one or more hazard event scenarios based on the characteristics of the multiple vulnerability features, The trained machine learning model determines, from the characteristics of the plurality of vulnerability features of the partition, a plurality of mitigation steps to reduce the damage tendency score for one or more hazard event scenarios, The trained machine learning model selects a proposed subset of one or more mitigation steps from the plurality of mitigation steps, For each subset of one or more mitigation steps, The trained machine learning model determines a corresponding updated damage tendency score based on a subset of the selected mitigation steps. Selecting one or more proposed subsets of mitigation steps, wherein the proposed subsets of mitigation steps correspond to the updated damage trend scores that result in a target reduction of the damage trend scores for the one or more hazard event scenarios. Selecting a proposed subset of one or more mitigation steps, To provide for display the proposed subset of one or more mitigation steps and the corresponding updated damage tendency score representation, Computer implementation methods, including those mentioned above.

2. The computer implementation method according to claim 1, wherein the imaging data includes one or more of the following: street view imaging data of the area, LIDAR data, high-resolution satellite image data, aerial image data, infrared image data, and user-provided images.

3. The computer implementation method according to claim 1, wherein the target reduction includes, for one or more subsets of mitigation steps, the maximum reduction between the updated damage tendency score and the damage tendency score.

4. The computer implementation method according to claim 1, wherein a first proposed subset of one or more mitigation steps for a first hazard scenario is different from a second proposed subset of one or more mitigation steps for a second hazard scenario.

5. Extracting the aforementioned multiple mitigation steps is The trained machine learning model includes identifying adjustments to the characteristics of the plurality of vulnerability features extracted from the imaging data, The computer implementation method according to claim 1, wherein determining the updated damage tendency score includes determining the updated damage tendency score based on the adjustments of the plurality of vulnerability features to the characteristics.

6. Selecting one or more subsets of mitigation steps from the plurality of mitigation steps includes, for each of at least two hazard event scenarios, selecting a corresponding proposed subset of one or more mitigation steps for each of the at least two hazard event scenarios. Providing the representations includes providing representations of the corresponding proposed subset for each of the at least two hazard event scenarios for display purposes. The computer implementation method according to claim 1.

7. Selecting one or more subsets of mitigation steps from the aforementioned plurality of mitigation steps, The computer implementation method according to claim 1, comprising selecting at least two different subsets of one or more mitigation steps from the plurality of mitigation steps.

8. A non-temporary computer storage medium encoded by a computer program, wherein the computer program includes an instruction that causes the data processing device to perform the operation described in any one of claims 1 to 7 when the data processing device is executed by the data processing device.

9. User devices and, One or more computers, and one or more storage devices that store instructions that, when executed by the one or more computers, are capable of causing the one or more computers to interact with the user device and perform the operations described in any one of claims 1 to 7. A system equipped with these features.

10. Receiving a request for one or more damage tendency scores for locations corresponding to a hazard event, wherein the location includes multiple sections; Receiving imaging data of the location including the plurality of sections, A machine learning model including multiple classifiers extracts the characteristics of multiple vulnerability features of each of the multiple sections from the imaging data, The machine learning model determines one or more damage tendency scores for the location, which represent a measure of the risk to the location caused by the hazard event, based on the characteristics of the plurality of vulnerability features. To provide a representation of one or more damage tendency scores for the location, including the plurality of sections corresponding to the hazard event, Computer implementation methods, including those mentioned above.

11. The computer implementation method according to claim 10, wherein the one or more damage tendency scores of the location include the individual damage tendency scores of each of the plurality of compartments of the location.

12. The computer implementation method according to claim 10, wherein the one or more damage tendency scores for the location include a global damage tendency score for the location.

13. The computer implementation method according to claim 10, wherein the location including the plurality of plots includes a neighborhood, a street including multiple houses, or a complex facility including multiple buildings.

14. The computer implementation method according to claim 10, wherein the hazard event is one of a wildfire, a flood, and an earthquake.

15. The computer implementation method according to claim 10, wherein receiving the request for one or more damage tendency scores includes receiving the GPS coordinates of the location.

16. The computer implementation method according to claim 10, wherein receiving the request further includes receiving a request for a mitigation step to reduce the hazard vulnerability of at least one of the plurality of compartments at the location, the mitigation step comprising a quantifiable measure to reduce the one or more damage tendency scores at the location.

17. The computer implementation method according to claim 10, wherein receiving imaging data includes receiving imaging data of the location including the plurality of parcels captured within a threshold time from the time of the request.

18. The machine learning model is trained on training data that includes multiple hazard events and multiple segments for each of the multiple hazard events, and the machine learning model determines one or more damage tendency scores from the characteristics of the multiple vulnerability features. The computer implementation method according to claim 10, comprising generating inferences between the characteristics of the plurality of vulnerability features of the plurality of sections at the location and the one or more damage tendency scores of the location using the machine learning model.

19. One or more non-temporary computer storage media encoded by a computer program, wherein the computer program includes instructions that cause one or more computers to perform the operations described in any one of claims 10 to 18 when executed by one or more computers.

20. One or more computers, and one or more storage devices that store instructions that, when executed by the one or more computers, are operable to cause the one or more computers to perform the operations described in any one of claims 10 to 18, A system equipped with these features.

Images