The temporal boundary of a wildfire

Abstract

To provide methods, systems and apparatus, including computer programs encoded on computer storage media, for generating a temporal range of a fire.SOLUTION: In some implementations, a server obtains a date when a fire occurred within a region. The server obtains satellite imagery of the region from before the date when the fire occurred. The server generates a first statistical distribution from the satellite imagery. The server determines a start date of the fire using the first statistical distribution. The server obtains second satellite imagery of the region from before and after the start date. The server selects a second set of imagery from the second satellite imagery from before the start date. The server generates a second statistical distribution from the second set of imagery. The server determines an end date of the fire using the second statistical distribution. The server provides the start date and the end date for output.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] This specification generally relates to computer simulations, and one particular implementation relates to generating a time range of wildfires using statistical distributions and satellite imagery.

Background Art

[0002] Wildfires can include forest fires, wildland fires, rural fires, etc., and can occur in various geographical regions. A wildfire can originate in one geographical area and spread to another geographical area over a period of time. Wildfires can be caused by factors such as climate, vegetation, and even human activities, to name a few.

Summary of the Invention

[0003] The techniques described in this specification generate a time range of historical wildfires across a geographical region. In particular, a server can generate a time range that includes a start date and an end date of a wildfire given its spatial boundaries in a geographical region, and the dates on which the wildfire occurred in the geographical region. The start date corresponds to the day the wildfire occurred. The end date corresponds to the day the wildfire ended.

[0004] Techniques for generating time ranges for wildfires are important because these ranges may be utilized and needed for other purposes. For example, a server could provide time ranges as training data for machine learning or another machine learning model pipeline to predict the duration of a fire. In another example, a server could provide the generated time ranges to a graphical user interface that can display the extent and duration of a wildfire. In yet another example, a server could provide the generated time ranges to a process attempting to identify the start date of a currently burning wildfire. In yet another example, an individual might want to analyze other sensory data before and after a wildfire occurs. Automatically detecting the start and end dates of a wildfire can help complete these investigations. For example, an analyst might need the time boundaries of a wildfire while attempting to estimate the annual number of tree deaths in a particular area caused by a wildfire.

[0005] In some implementations, the server can receive the date on which a wildfire occurred within a geographical area. This date could be the start date, the end date, or any date between the start and end dates. The geographical area can be defined, for example, by location data that includes longitude and latitude coordinates and the area or region surrounding these coordinates.

[0006] The server can communicate with a satellite image database to retrieve satellite imagery from various geographical areas. The satellite image database can store satellite imagery, including infrared and other satellite data, from various geographical areas of the Earth over different time periods. The satellite imagery may also include other processed satellite imagery products, such as MODIS MCD45A1, which includes pixel-by-pixel burning and quality assurance information. The satellite image database can index satellite imagery by several factors, such as location data, time, date, and geographical area boundaries.

[0007] The server can identify and retrieve satellite imagery using location data associated with a geographical area and the input date on which the fire occurred. Specifically, the server can provide location data and the input date to a satellite imagery database and receive satellite imagery from periods prior to the provided input date. In some implementations, the server can provide location data and the input date to a satellite imagery database and receive satellite imagery from periods prior to and after the provided input date.

[0008] In some implementations, the server can generate a statistical distribution based on satellite imagery acquired for a specific geographical area from a period prior to the input date. As further described below, the server can generate a baseline of satellite imagery corresponding to the period prior to the input date, where each image within the baseline represents a geographical area over a defined period. Once the server generates the statistical distribution, it can identify the start date of the wildfire.

[0009] In some implementations, the server can determine the start date of a wildfire by comparing satellite imagery with a statistical distribution. The server can retrieve satellite imagery corresponding to a geographical area period prior to the input date, for example, a geographical area period three months prior to the input date. Three months prior to the input date is chosen to ensure that the statistical distribution does not contain or include any fire pixels. Furthermore, the server can filter the satellite imagery within the smallest known periphery of the wildfire. This is further explained below. The server can identify the start date of a wildfire by comparing satellite imagery with a statistical distribution.

[0010] In some implementations, the server can determine the end date of a wildfire by generating an additional statistical distribution. The server can generate a second statistical distribution based on satellite imagery acquired for a specific geographic area from a period prior to the determined start date. For example, the server filters the acquired satellite imagery within the most recent known periphery of the wildfire within the geographic area. As further described below, the server can improve the accuracy of identifying the end date by generating an additional statistical distribution using the determined start date rather than the provided input date. When generating the second statistical distribution, the server can generate another baseline of satellite imagery corresponding to the period prior to the determined start date, where each image in the baseline represents a geographic area over a defined period. Once the server generates the second statistical distribution, it can identify the end date of the wildfire.

[0011] In some implementations, the server can determine the end date of a wildfire by comparing satellite imagery with a second statistical distribution. The server can then retrieve satellite imagery corresponding to a period in a geographical area after the determined start date. Selecting satellite imagery after the determined start date is necessary to determine the end date, while selecting satellite imagery before the provided input date is necessary to determine the start date.

[0012] The server can then provide, for output, a time range for the wildfire, such as the determined start and end dates of the wildfire. The output may be for connected client devices, displays, or one or more other applications.

[0013] In one common embodiment, the method is performed by a server. The method involves the server obtaining the date on which a fire occurred within a geographical area; the server obtaining first satellite imagery of the geographical area prior to the date on which the fire occurred; the server selecting a first set of images from the first satellite imagery prior to the date on which the fire occurred; the server generating a first statistical distribution from the first set of images; the server determining the start date of the fire based on a comparison between the first satellite imagery and the first statistical distribution; and the server obtaining second satellite imagery of the geographical area prior to and after the determined start date on which the fire occurred. The process includes: acquiring images; the server selecting a second set of images from a second set of satellite images prior to the start date of the fire within the geographical area, wherein the amount of the second set of images is less than the amount of the first set of images; the server generating a second statistical distribution from the second set of images prior to the start date of the fire within the geographical area; the server determining the end date of the fire based on a comparison between the second set of satellite images occurring after the start date and the second statistical distribution; and providing a range for output by the server that includes the start date and end date of the fire within the geographical area.

[0014] Other embodiments and other aspects of the present disclosure include corresponding systems, devices, and computer programs encoded on computer storage devices and configured to perform actions of this method. One or more computer systems may be configured by software, firmware, hardware, or a combination thereof installed on the system that causes the system to perform actions during operation. One or more computer programs may be configured by having features that, when executed by a data processing device, cause the device to perform actions.

[0015] Each of the above and other embodiments may optionally include one or more of the following features individually or in combination. For example, one embodiment may include all of the following features in combination.

[0016] In some implementations, the method includes obtaining the date a fire occurred within a geographical area by the server obtaining a location that describes the geographical area, and the server obtaining the date a fire occurred within the geographical area, wherein the date may correspond to the start date of the fire, the end date of the fire, or another date within the range.

[0017] In some implementations, the method includes obtaining a first satellite image of a geographic area prior to the date on which a fire occurred within that geographic area, by providing a satellite image database with a location describing the geographic area where the fire occurred and the date on which the fire occurred, and by obtaining a first satellite image showing the geographic area from a period prior to the date in response to the provision of the location and date.

[0018] In some implementations, the method includes: selecting a first set of images from first satellite imagery prior to the date on which a fire occurred within a geographic area; identifying a period for generating a first statistical distribution, wherein the period corresponds to a period prior to the date; selecting a first set of images from first satellite imagery based on the period; determining whether one or more pixels from the selected first set of images indicate a fire; and adjusting the selected first set of images by comparing one or more pixels indicating a fire with a threshold.

[0019] In some implementations, the method generates a first statistical distribution from a first set of images, and in response to adjusting a first set of selected images, the server generates one or more parameters of the first statistical distribution by: the server determining the number of pixels indicating fire in each day of the first set of images; the server determining the average number of pixels indicating fire in each day of the first set of images; and the server generating the first statistical distribution by setting the determined average to one or more parameters.

[0020] In some implementations, the method includes determining the start date of a fire based on a comparison of a first satellite image with a first statistical distribution, by a server identifying satellite images from the first satellite image corresponding to a day prior to the date the fire occurred, by a server determining the number of pixels that appear to indicate a fire from the identified satellite images for that day, by a server generating a likelihood that the satellite images from the first satellite image indicate a fire based on the first statistical distribution and the number of pixels determined to indicate a fire, by a server comparing the likelihood to a threshold, and in response to determining that the likelihood does not exceed the threshold, by a server obtaining another satellite image from the first satellite image corresponding to a different day prior to the date the fire occurred, in order to determine the start date.

[0021] In some implementations, the method further includes, in response to determining that the likelihood for a particular day exceeds a threshold, the method having the server acquire additional satellite imagery for a predetermined number of days prior to a particular day; for each of the predetermined days, the server determining the number of active fire pixels from the additional satellite imagery for that day; the server generating a likelihood that the additional satellite imagery for that day does not indicate a fire, based on a first statistical distribution and the number of pixels determined to appear to indicate a fire; and, in response to determining that each of the predetermined days does not indicate a fire, the server determining that the day corresponding to the identified satellite imagery corresponds to the start date of the fire.

[0022] In some implementations, the method includes, by providing a server with a location describing the geographic area where the fire occurred and the start date of the fire to a satellite image database, and by obtaining a second satellite image showing the geographic area from the period before and after the start date, in response to providing the location and start date.

[0023] In some implementations, the method includes: selecting a second set of images from a second set of satellite imagery prior to the start date of a fire occurring within a geographical area, wherein the quantity of the second set of images is less than the quantity of the first set of images; identifying a period for generating a second statistical distribution, wherein the period corresponds to a period prior to the start date; selecting a second set of images from the second set of satellite imagery based on the period; determining whether one or more pixels from the selected second set of images indicate a fire; and adjusting the selected second set of images by comparing one or more pixels indicating a fire with a threshold.

[0024] In some implementations, the method generates a second statistical distribution from a second set of images prior to the start date of a fire within a geographical area, and in response to adjusting the second set of selected images, the server generates one or more parameters of the second statistical distribution by determining the number of pixels indicating a fire in each day of the second set of images, determining the average number of pixels indicating a fire in each day of the second set of images, and generating the second statistical distribution by setting the determined average to one or more parameters.

[0025] In some implementations, the method determines the end date of a fire based on a comparison of a second satellite image occurring after the start date with a second statistical distribution, which includes: the server identifying satellite images from the second satellite image corresponding to a date after the start date; the server determining the number of pixels that appear to indicate a fire from the identified satellite image for that day; the server generating a likelihood that the satellite image from the second satellite image contains indications of a fire, based on the second statistical distribution and the number of pixels determined to indicate a fire; the server comparing the likelihood to a threshold; and, in response to determining that the likelihood exceeds the threshold, the server obtaining another satellite image from the first satellite image corresponding to another day after the date the fire occurred, in order to determine the end date.

[0026] In some implementations, in response to determining that the likelihood for a particular day exceeds a threshold, the method includes the server obtaining additional satellite images for a predetermined number of days after the particular day, determining, for each day of the predetermined number of days, the number of pixels from the additional satellite image for that day by the server, generating, by the server, a likelihood that the additional satellite image for that day does not contain an indication based on a second statistical distribution and the determined number of pixels that appear to indicate a fire, and determining, by the server, that the day corresponding to the identified satellite image corresponds to an end day of a fire in response to determining that each day of the predetermined number of days does not indicate a fire.

[0027] In some implementations, the method includes the first and second statistical distributions being different Poisson distributions.

[0028] Details of one or more embodiments of the subject matter of this specification are set forth in the accompanying drawings and the description below. Other features, aspects, and advantages of the subject matter will become apparent from the description, the drawings, and the claims.

Brief Description of the Drawings

[0029] [Figure 1] FIG. is a block diagram illustrating an example of a system for generating a time range of a wildfire. [Figure 2A] FIG. is a block diagram illustrating an example of a system for receiving satellite images and generating a time range of a wildfire from the satellite images. [Figure 2B] FIG. is a block diagram illustrating an example of a system for generating a statistical distribution when determining a time range of a wildfire. [Figure 3] FIG. is a flowchart illustrating an example of a process for generating a time range of a wildfire.

[0030] Similar reference numbers and names in various drawings refer to similar elements. The components shown herein, their connections and relationships, and their functions are illustrative and not intended to limit the implementations described and / or claimed herein. [Modes for carrying out the invention]

[0031] Figure 1 is a block diagram showing an example of a system 100 for generating a time range for wildfires. System 100 includes an image server system 102, a satellite image database 104, a time database 106, and a spatial database 108. System 100 also includes a monitor 126 connected to the image server system 102 to display and allow a user to interact with the image server system 102. In short, system 100 can generate a time range for wildfires identified from the date the wildfire occurred and the corresponding location of the wildfire. System 100 can generate multiple statistical distributions and compare acquired satellite images with the generated statistical distributions to identify the time range of the wildfire, e.g., the start date and end date.

[0032] System 100 attempts to generate a time range for wildfires, as the start and end dates of wildfires are often missing from the database storing fire information. By generating and storing this information, System 100 can add to and improve the overall modeling and monitoring of wildfires. This information can also be useful for various wildfire applications, as described below.

[0033] In some implementations, the image server system 102 may include one or more servers or computers connected locally or via a network. System 100 may include a network that could be, for example, a local network, a Wi-Fi network, an intranet, an internet connection, a Bluetooth connection, or any other connection that enables the image server system 102 to communicate with various databases and various computers, for example, to send and receive data. Figure 1 shows various operations in stages (A) to (J) that may be performed in the order shown or in a different order.

[0034] The image server system 102 can also communicate with satellites via a satellite network. Satellites can capture media, such as images and videos, from the Earth's geographical regions. The image server system 102 can communicate with satellites to request and receive captured media from various geographical regions of the Earth. This will be further explained below.

[0035] In some implementations, the satellite image database 104 may include one or more databases that store satellite imagery representing one or more geographical regions of the Earth. The satellite imagery may include high-quality satellite imagery, medium-quality satellite imagery, noisy satellite imagery, and other types of satellite imagery. Furthermore, the satellite image database 104 may store satellite imagery representing one or more geographical regions of the Earth over a period of time. For example, the satellite imagery may include media, such as images and videos, of a specific geographical region of the Earth over the past year, five years, ten years, fifteen years, or longer, from the current time. The satellite image database 104 may include satellite imagery for each day over a historical period. In some cases, the satellite image database 104 may include multiple images or videos from each day of a geographical region.

[0036] The satellite image database 104 can store satellite media from multiple geographical regions of the Earth over various periods. For example, the satellite image database 104 can store satellite images from geographical region 1 to geographical region N, and for each of those geographical regions, it can store satellite images for various periods. In these examples, the satellite image database 104 can store satellite images for each day over the past 10 years for geographical region 1, and satellite images for each day over the past 20 years for geographical region 2.

[0037] In some implementations, the dimensions of satellite imagery may correspond to the camera used on the satellite and the satellite's location. Satellites can orbit the Earth at a specific distance and capture images of the Earth from that distance. For example, a satellite can orbit the Earth in low orbit at a distance of 800 kilometers (km) or 497.097 miles. In some implementations, the camera associated with the satellite may have zoom capabilities.

[0038] In some implementations, the resolution of a satellite camera may be set to a standard resolution to ensure uniformity across all satellite images. For example, the camera resolution may be set to 375 meters horizontally and vertically when capturing a geographical area of ​​the Earth. In other implementations, different satellites may have different resolution capabilities. Therefore, satellite cameras may adjust their zoom capabilities to ensure uniform resolution across satellite images from each satellite.

[0039] In some implementations, the time database 106 may include one or more databases that store periods of satellite imagery captured by satellites and stored in the satellite imagery database 104. These periods may include, for example, defined dates, date ranges, and time ranges. For instance, the time database 106 may store a time range from January 1, 2018 to January 1, 2019, or from 12:00 PM on January 1, 2018 to 12:01 PM on January 1, 2019.

[0040] Furthermore, the time database 106 can store even more specific date ranges. More specific date ranges may include specificity down to the hour and minute ranges. For example, the time database 106 can store the time range from January 1, 2018, 12:00 PM to January 1, 2019, 12:30 PM.

[0041] In some implementations, a satellite can provide satellite imagery to a satellite image database 104 along with metadata. The metadata may include location data for the geographical area where the satellite imagery was captured, data identifying the satellite from which the imagery was captured, the date and time the imagery was captured, and location data indicating where the satellite was located in space when the imagery was captured. For example, if a satellite captured an image of the geographical area of ​​San Diego, the satellite could associate the following metadata with the image: (1) location data for the geographical area as 32.7515, -117.1364, (2) date 12:05 PM PT on 19 May 2019 when the satellite imagery was captured, (3) data indicating the name and model of the satellite from which the imagery was captured, e.g., satellite 111 and model 1, and (4) location data of the satellite at the time the imagery was captured, such as 800 km above the location coordinates 32.7515, -117.1364. The satellite image database 104 can identify the requested satellite image by comparing the date and location of the wildfire provided by the image server system 102 with the metadata of the satellite image.

[0042] The image server system 102 can accordingly distribute metadata to each of the databases within system 100. For example, when a satellite provides satellite imagery to the image server system 102, the image server system 102 can provide the satellite imagery and metadata to the satellite image database 104 for storage purposes. In addition, the image server system 102 can extract time data from the metadata and provide the time data to the time database 106. The time data can indicate the date and time the satellite imagery was captured by the satellite. Furthermore, the image server system 102 can extract geographical location data from the metadata and provide the location data to the spatial database 108. The image server system 102 can perform this extraction and distribution process for each satellite imagery acquired from the satellite.

[0043] In some implementations, the spatial database 108 may include one or more databases that store location data for locations identified in satellite imagery captured by satellites and stored in the satellite imagery database 104. The location data may include, for example, latitude and longitude coordinates, addresses, landmark names, dates that identify cities, states, and counties, and other data that identify locations on Earth.

[0044] The spatial database 108 may include one or more bounding boxes that indicate the location of an area where a wildfire started, for example, a geographical region or coordinates. A bounding box may include, for example, a polygon that indicates the area around the geographical region where the wildfire began or where the fire started. The bounding boxes may be applied to satellite imagery so that the image server system 102 can determine where to count fire pixels. For example, as will be further described below, the image server system 102 can count the number of fire pixels within the bounding boxes applied to the satellite imagery.

[0045] A polygon may contain one or more geographical features and one or more wildfire features. For example, one or more wildfire features may include one or more areas where no fire is occurring, one or more areas where a fire is actively burning, one or more burning areas where a fire recently occurred, and one or more scar areas resulting from the fire. The spatial database 108 may be populated by the image server system 102 or another external service.

[0046] The image server system 102 can store data in the time database 106 and the spatial database 108 in response to acquiring satellite imagery. In some implementations, the image server system 102 can store time and spatial data from the satellite imagery metadata to ensure that the satellite imagery database 104 can provide appropriate satellite imagery when requested. In this case, if the image server system 102 requests satellite imagery outside the data contained in the time database 106 and the spatial database 108, the satellite imagery database 104 may return an error message indicating that the requested data is outside the satellite imagery range.

[0047] In some implementations, an external process to the image server system 102 can perform population operations on the satellite image database 104, the time database 106, and the spatial database 108. In this case, the image server system 102 can perform processes related to determining the time range of a wildfire by accessing the time database 106, the spatial database 108, and the satellite image database 104.

[0048] During stage (A), the image server system 102 can receive an input date 110 in which the wildfire occurred. As shown in system 100, the input date 110 corresponds to July 5, 2019. In some implementations, an external process can access the time database 106 and provide the input date 110 to the image server system 102. In some implementations, a user interacting with the image server system 102 can request a time range for the wildfire and provide the input date 110 to the image server system 102. In addition, input data may be provided by, for example, a fire metadata database. The input date 110 can indicate the date the wildfire occurred, and in addition, it can provide the image server system 102 with an instruction that the satellite image database 104 contains satellite images of the wildfire at the corresponding geographical location. In other implementations, the image server system 102 can provide the time database 106 with a request to return the date on which the wildfire occurred at a specific geographical location.

[0049] During stage (B), the image server system 102 can receive the location 112 where the wildfire occurred, based on the input date 110 of the wildfire. As shown in system 100, the wildfire location may correspond to location coordinates such as 33.8121N, -117.91899E. Location 112 may also include other descriptors, such as the name of a landmark, the name of a city, the name of a geographical area, or a specific address.

[0050] In some implementations, the image server system 102 can retrieve polygons from the spatial database 108 that indicate the location where a wildfire first started. The image server system 102 can acquire polygons based on the input date 110 and location 112. To limit the amount of processing performed by the image server system 102, the image server system 102 can apply the acquired polygons to the acquired satellite images after retrieving the satellite images from the satellite image database 104. In this case, the polygons can spatially limit the number of pixels searched by the image server system 102 within the satellite images.

[0051] Even if the input date 110 is actually the start or end date of the wildfire, the input date 110 does not necessarily indicate the start or end date of the wildfire to the imager server system 102. Rather, the input date 110 may correspond to another date, such as a day between the start and end dates, when the wildfire was active or ongoing. The image server system 101 can determine where the input date 110 falls within the time range of the wildfire by continuing the process, as will be further explained below.

[0052] During stage (C), the image server system 102 can provide the satellite image database 104 with the input date 110 and the location 112 where the wildfire occurred. By providing the input date 110 and location 112 to the satellite image database 104, the image server system 102 can obtain satellite images for locations and time periods prior to the input date 110. In some implementations, the image server system 102 can provide the satellite image database 104 with a period prior to the input date 110 in order to obtain satellite images. The period may include, for example, one year prior to the input date 110, two years prior to the input date 110, five years prior to the input date 110, or any other period prior to the input date 110 that is sufficient to construct a statistical distribution. For example, the image server system 102 may need satellite images for at least one year prior to the input date 110 in order to construct a statistical distribution, and as a result, it may request satellite images from the satellite image database 104 that are more than two years prior to the input date 110.

[0053] In some implementations, the satellite image database 104 can index its stored satellite images by location. For example, the image server system 102 can also provide locations 112 to the satellite image database 104, and the satellite image database 104 can access satellite images that include locations 112 within its field of view, such as location coordinates or other geographical or locational descriptions. The satellite image database 104 can determine the location 112 within the field of view of the corresponding satellite image by analyzing the metadata of the satellite image.

[0054] Next, the satellite image database 104 can filter images containing location 112 by time range. For example, the satellite image database 104 can exclude or filter satellite images that have a timestamp and date after the input date 110 and exist outside the period before the input date 110. For example, the image server system 102 can instruct the satellite image database 104 to provide all satellite images showing location 112 from two years prior to the date of July 5, 2019, for example, from July 5, 2017 to July 5, 2019. The image server system 102 can exclude all satellite images showing location 112 before the date of July 5, 2017 and after the date of July 5, 2019.

[0055] The satellite image database 104 can identify and provide satellite imagery 114 that meets the criteria of the image server system 102 and can be returned to the image server system 102. As shown in system 100, satellite imagery 114 may include one or more images or videos of satellite imagery including location 112 and may have timestamps that fall within a time range from two years prior to input date 110 to input date 110. In some examples, satellite imagery 114 may include satellite imagery for each day from July 5, 2017 to July 5, 2019, or may include multiple satellite imagery or videos for each day from July 5, 2017 to July 5, 2019. In some cases, one or more days between July 5, 2017 and July 5, 2019 may not include satellite imagery if the satellite did not capture images around location 112 on those days.

[0056] In some implementations, each media of the satellite image 114 can have a specified resolution. For example, as described above, each media can have a resolution of 325 meters horizontally and 325 meters vertically. This resolution allows the image server system 102 to visually inspect not only location 112 but also the surrounding area of ​​location 112. Since wildfires can originate in a specific location and move to another, by viewing a wide area adjacent to location 112, such as within a bounding box around location 112, the image server system 102 can improve its determination of the start and end dates of the wildfire. For example, a bounding box or polygon can correspond to an area in the satellite image that indicates the area encompassing the wildfire. The bounding box can include the smallest area covering the wildfire in the satellite image, or an area encompassing both the wildfire and the area outside of it.

[0057] In addition, if location 112 lies on the edge of satellite image 114, the satellite image database 104 can also provide satellite imagery adjacent to location 112 within the edged satellite imagery. In this case, the image server system 102 can ensure that location 112 is always enclosed by an area of ​​the image, such as the image resolution, even if location 112 shown in the satellite imagery lies on, for example, a vertical or horizontal edge of the satellite imagery. In some implementations, the image server system 102 can indicate a boundary region or area around location 112 to the satellite image database 104. Then, once the satellite image database 104 identifies satellite imagery that meets the criteria of the image server system 102, the satellite image database 104 can identify additional satellite imagery to be provided if the initially identified satellite imagery does not fall within the boundary region or area criteria.

[0058] In some implementations, the satellite image database 104 can provide identified satellite images 114 to the image server system 102 via a network. In other implementations, the satellite image database 104 can provide the image server system 102 with an index of the identified satellite images 114 in order to perform retrieval of the satellite images 114. In some cases, the satellite image database 104 can also provide a link to the identified satellite images 114 for rapid retrieval by the image server system 102. The link may include, for example, a zip file, access to cloud storage, or some other form for downloading the satellite images 114.

[0059] During stage (D), the image server system 102 can generate a statistical distribution based on the identified satellite imagery 114. In some implementations, the image server system 102 can generate a statistical distribution for comparison with satellite imagery to determine the likelihood of wildfires in the satellite imagery. In particular, the statistical distribution corresponds to a frequency model of satellite noise in a specific geographical area over a certain period of time. In some examples, the statistical distribution can be modeled using a Poisson, Gaussian, or normal distribution.

[0060] In some implementations, the satellite image 114 can show various landmarks, buildings, roads, and other geographical features. Other geographical features may include, for example, rivers, seas, lakes, hills, and plains. The satellite image 114 can also show people, vehicles, animals, and other features commonly found in the geographical area. The satellite image 114 may also include noise characteristics such as glare, flashes, and distorted pixels, which may make it difficult for the image server system 102 to distinguish wildfires from noise.

[0061] In some implementations, satellite imagery can show various fire areas. These areas may include, for example, one or more areas where there is no fire, one or more areas that are actively burning, one or more burning areas where a fire recently occurred, and one or more scar areas resulting from a fire. Areas where there is no fire may indicate another geographic area where there is no fire. Areas that are actively burning may include, for example, areas showing active fire, smoke, or embers. Burning areas where a fire recently occurred may indicate geographic areas where a fire once flared up but is no longer active, but these geographic areas are still hot and dangerous. Scar areas may indicate geographic areas where a fire once flared up but is no longer active, and these geographic areas are no longer hot or dangerous.

[0062] The image server system 102 can generate and utilize statistical distributions to improve the accuracy of detecting ongoing wildfires in satellite imagery. Specifically, the image server system 102 can begin by first identifying the number of pixels in satellite imagery 114 that appear to indicate wildfires, and then compare the satellite imagery with the generated statistical distribution. Some of the identified pixels may not indicate wildfires, but rather be noisy pixels. To mitigate this problem, the image server system 102 can compare the number of identified pixels in satellite imagery 114 that appear to indicate wildfires with the generated statistical distribution and filter out images with noisy characteristics.

[0063] To enable such comparisons, the image server system 102 can construct a statistical distribution using historical satellite imagery obtained from a satellite image database 104 that shows the same locations as location 112 that does not show an ongoing wildfire. For example, the image server system 102 must first identify the period prior to the input date 110 within the acquired satellite imagery 114 that does not show a wildfire. In addition to identifying a date three months prior to the input date 110 to ensure that no wildfires are found in the corresponding satellite imagery, the image server system 102 can identify historical weather forecasts and other meteorological databases to confirm whether a fire existed in a specific geographic area, such as a geographic area identified in the satellite imagery. By generating a satellite image baseline, the image server system 102 can identify any pixel in the satellite imagery that lies outside the baseline norm and indicate that the corresponding satellite imagery appears to show a wildfire. In this case, the image server system 102 first determines a date three months prior to the input date 110. In the example of system 100, the date three months prior to input date 110 is April 5, 2019.

[0064] The image server system 102 can determine a date three months prior to the input date 110, since wildfires generally do not last longer than three months. By identifying a date three months prior to the input date 110, the image server system 102 can safely assume that the same wildfire identified by input date 110 and location 112 is inactive or not burning in the corresponding satellite image, for example, the satellite image for a date three months prior to input date 110. In some implementations, if the image server system 102 identifies a wildfire still occurring in satellite image 114 on a date three months prior to input date 110, the image server system 102 can identify an even earlier date. In some cases, the image server system 102 can identify whether one or more other wildfires were active during the period between input date 110 and a date three months prior to input date 110. One or more other wildfires may have different origins and correspond to wildfires that did not cause the wildfire identified by input date 110. The image server system 102 can access one or more public wildfire databases to determine whether another wildfire was active during this time window. However, it is rare for another wildfire to be active within the same geographical area as the wildfire identified by input date 110.

[0065] For example, the image server system 102 may set the date back one month prior to April 5, 2019, for example, March 5, 2019. If the image server system 102 determines that there are no pixels in the satellite image 114 that appear to indicate a wildfire on March 5, 2019, the image server system 102 may set March 5, 2019, as the end date of the time range for constructing the statistical distribution. The image server system 102 can iteratively perform the process of identifying the end date of the time range, for example, by going back one day, one month, or one week until a start date without a wildfire is shown in the corresponding satellite image. In this case, the image server system 102 cannot set this determined end date as the actual start date of the wildfire because the wildfire may not actually be active for, for example, one day, one week, or one month after the end date.

[0066] In some implementations, the image server system 102 can determine a date one year prior to the determined end date for a time range. This date one year prior to the determined end date for the time range is known as the start date of the time range. The image server system 102 can identify the start and end dates of a time range for identifying satellite images and constructing a statistical distribution. Continuing with the example of system 100, the image server system 102 can determine that the start date of the time range is April 5, 2018, one year prior to the end date, for example, April 5, 2019.

[0067] However, in order for the image server system 102 to set the start date to April 5, 2018, the image server system 102 may use data from an external database of historical fires to ensure that the satellite imagery showing location 112 does not show active wildfires between the start date (e.g., April 5, 2018) and the end date (e.g., April 5, 2019). If the image server system 102 identifies data from the external database of historical fires indicating, for example, that wildfires existed or were active in a geographical area from April 5, 2018 to April 5, 2019, the image server system 102 may adjust the start or end date accordingly. The manner in which the image server system 102 detects or identifies ongoing wildfires in satellite imagery is described further below.

[0068] To determine whether satellite imagery within a given time range indicates an ongoing wildfire, the image server system 102 can identify the number of fire pixels in the satellite imagery and compare the number of fire pixels to a threshold. For example, the image server system 102 can run a fire detection algorithm on the satellite imagery to determine how many pixels contain fire. The fire detection algorithm could include, for example, an active fire detection algorithm. If the number of fire pixels in the satellite imagery is less than the threshold, the image server system 102 can indicate that a particular satellite imagery does not contain fire.

[0069] For example, the image server system 102 can select a subset of satellite images from satellite imagery 114, including both endpoints, from April 5, 2018 to April 5, 2019. The image server system 102 can process each image or video from each day within that time range and search for pixels that appear to indicate a wildfire. If the image server system 102 detects pixels that appear to indicate a wildfire on April 8, 2018, for example, if the number of pixels is greater than a threshold and no pixels are detected on other days, the image server system 102 can adjust the time range to April 9, 2018 to April 5, 2019. Alternatively, if the image server system 102 detects pixels that appear to indicate a wildfire on April 1, 2019, for example, if the number of pixels is greater than a threshold, the image server system 102 can adjust the time range to April 5, 2018 to March 31, 2019.

[0070] In some cases, if the image server system 102 detects pixels that appear to indicate a wildfire in the middle of a time range, such as December 1, 2018 or November 1, 2018, for example, if the number of pixels is greater than a threshold, the image server system 102 may exclude those days from the time range and adjust the time range into multiple consecutive sub-time ranges. For example, a consecutive sub-time range could include April 5, 2018 to October 31, 2018, November 2, 2018 to November 31, 2018, and December 2, 2018 to April 5, 2019. Thus, the image server system 102 can create a single consecutive time range or multiple consecutive sub-time ranges to construct a satellite image distribution.

[0071] In some implementations, the image server system 102 can determine a date nine months prior to a date three months prior to the input date 110. For example, if the image server system 102 determines that satellite images dated April 5, 2019, do not appear to show wildfires, the image server system 102 can identify a date nine months prior to April 5, 2019, as the start of the time range. The end date of the time range would be April 5, 2019. Nine months prior to April 5, 2019, corresponds to July 5, 2018. Therefore, the image server system 102 can construct a statistical distribution based on the time range from July 5, 2018 to April 5, 2019.

[0072] If the image server system 102 bases its statistical distribution on a shorter 9-month time range, for example, July 5, 2018 to April 5, 2019, it can generate the statistical distribution more quickly and save processing speed / power. However, if the image server system 102 bases its statistical distribution on a longer 1-year time range, for example, April 5, 2018 to April 5, 2019, it can generate a more accurate and representative statistical distribution for location 112. In some examples, wildfires may typically occur in a specific area, for example, the California region, during a standard 3-month fire season that begins around August and September. For fires occurring during this fire season, a 9-month window can be chosen instead of a 1-year window to reduce the likelihood that satellite imagery that incidentally includes wildfires from the previous year's fire season will be selected in the baseline distribution. In some implementations, the image server system 102 can remove fire seasons corresponding to a time range from satellite imagery to reduce the likelihood that fires are accidentally included within a time range or baseline distribution.

[0073] In response to identifying satellite images within a time range identified from satellite image 114, the image server system 102 can determine the number of pixels in the identified satellite images that appear to indicate wildfires for each day. For example, the image server system 102 can analyze satellite images for each day within the identified time range from April 5, 2018 to April 5, 2019, and determine the number of fire pixels for each day. In some implementations, the image server system 102 can apply polygons retrieved from the spatial database 108 to the identified satellite images. By cropping the identified satellite images into polygons, the image server system 102 can reduce the amount of pixels searched across the identified satellite images. In this case, the image server system 102 can search across pixels in the identified satellite images within the retrieved polygons.

[0074] The image server system 102 can, for example, count 10 fire pixels on April 5, 2018, count 11 fire pixels on April 6, and continue counting satellite images each day until April 5, 2019. The same process applies when analyzing satellite images over a one-year time range, or a time range having multiple consecutive sub-time ranges.

[0075] In some implementations, the image server system 102 can normalize the number of fire pixels detected each day. For example, the image server system 102 can analyze one satellite image on April 5, 2018, 50 satellite images on April 6, 2018, and 10 satellite images on April 7, 2018. The image server system 102 can normalize the detections by dividing by the average number of fire pixels detected each day. The reason for normalization is to ensure that the generated statistical distribution produces a smooth curve. If the values ​​are not normalized, the curve of the generated distribution may contain jagged edges, which can distort probability predictions.

[0076] In some implementations, after determining the number of fire pixels each day within a satellite image or within a satellite image polygon for an identified time range, the image server system 102 can generate a statistical distribution. For example, the image server system 102 can generate a frequency histogram of the number of identified pixels in a satellite image that appear to indicate an ongoing wildfire. The frequency histogram can be shown graphically with the number of days on the X-axis and the number of fire pixels on the Y-axis.

[0077] Next, the image server system 102 can determine the mean of the frequency histogram and fit other data descriptors, such as the mean or standard deviation, to a statistical distribution. For example, the image server system 102 can fit the mean of the frequency histogram to a Poisson distribution or another statistical distribution. In a Poisson distribution, the mean is equivalent to the variance. In system 100, the average number of pixels identified to represent wildfires over a given time interval, for example, one year, is equivalent to the variance of a Poisson distribution.

[0078] In some implementations, if a significant number of wildfire events are unintentionally included in the baseline distribution, the baseline distribution may no longer have equivalent means and variances. In other words, the baseline distribution may no longer follow a Poisson distribution. In this case, the image server system 102 can perform a test to determine whether the mean and variance of the baseline distribution, for example, the statistical distribution, are equivalent. The test is performed, for example, by analyzing the mean of the sample in the statistical distribution and analyzing the variance of the sample. If the image server system 102 detects that the mean and variance of the statistical distribution are not equivalent, the image server system 102 can adjust the time range and the corresponding satellite imagery and construct a new statistical distribution based on the satellite imagery from the adjusted time range. The image server system 102 can repeat this process until a statistical distribution with equivalent means and variances is constructed. Once the statistical distribution is constructed, only one parameter, such as the average number of pixels that appear to indicate a wildfire, is needed to determine the probability of the event. The following equation shows the probability distribution function of a Poisson random variable.

[0079]

number

[0080] In Equation 1 above, which represents the Poisson distribution, the value λ corresponds to the average number of pixels that appear to indicate a wildfire from the baseline. The value k corresponds to the number of times a random variable or event occurs, for example, the number of pixels that appear to indicate a wildfire from the input image. The resulting value, f(k;λ), corresponds to the probability representing the Poisson probability distribution function. In some implementations, the image server system 102 can return the Poisson distribution 116 once the value λ is determined.

[0081] The Poisson distribution 116 can provide a resulting value during comparison with the input image, which indicates the likelihood that the event will occur within a given period. In particular, the probability density function (PDF) of the Poisson distribution 116 can return the probability that a particular day with k counted pixels occurred naturally from the noise distribution. Therefore, a lower value output by the PDF indicates a higher probability that an external force, such as a wildfire, contributed to the k count. Conversely, a higher value output by the PDF indicates a lower probability that an external force contributed to the k count, and that the image is more closely similar to an image from the baseline distribution. For example, if the probability corresponds to a low value such as 0.01, the image server system 102 can determine that the image is likely to show a wildfire. Conversely, if the probability corresponds to a high value such as 0.90, the image server system 102 can determine that the image is likely not to show an ongoing wildfire.

[0082] During stage (E), the image server system 102 can determine the start date of the wildfire based on the generated statistical distribution and satellite imagery 114. For example, as shown in system 100, the image server system 102 can generate a statistical distribution 116 during stage (D). The image server system 102 then compares each of the satellite imagery 114, starting from the input date 110 of July 5, 2019, with the generated statistical distribution 116.

[0083] First, the image server system 102 determines from the satellite image 114 the number of pixels that appear to indicate wildfires in the satellite image from July 5, 2019. In some implementations, the image server system 102 can use a fire detection algorithm to determine whether a pixel indicates a wildfire. In other implementations, the image server system 102 can determine whether a pixel indicates a wildfire by analyzing the brightness of the pixels. For example, the image server system 102 can determine the number of pixels that appear to indicate a wildfire in the satellite image from July 5, 2019 by analyzing the brightness of each pixel and comparing the brightness of the pixels to a threshold. For example, the image server system 102 can determine the brightness of a pixel by averaging the red, blue, and green (RGB) values ​​of the pixels. In some implementations, the image server system 102 can determine which pixels are on fire by analyzing wavelengths outside the RGB values. For example, infrared wavelengths may be outside the RGB wavelengths. The lower the average, the lower the brightness of the pixel, and vice versa. In another example, the image server system 102 can determine the brightness of a pixel by calculating the relative brightness value using the following formula. Y=0.2126*R+0.7152*G+0.0722*B (2)

[0084] Equation 2 above shows that the relative brightness value of a pixel is calculated based on the pixel's RGB values. For example, if a pixel's red value is 100, green value is 100, and blue value is 100, the relative brightness value is 100. A higher relative brightness value indicates a higher brightness of the pixel. While pixel brightness does not indicate whether a pixel is a wildfire or noise, it can be an indication that the pixel appears to be on fire. In some implementations, the image server system 102 can also analyze the intensity of pixels to determine whether a wildfire is present within a pixel. In other implementations, the image server system 102 can rely on an external source to determine whether a corresponding pixel is a wildfire. For example, the image server system 102 can use the VIIRS (Visible Infrared Imaging Radiometer Suite) Active Fire Product developed by NASA to combine the infrared band from satellite imagery with the pixel's RGB band to determine whether a pixel was on fire. Typically, each satellite includes its own detection algorithm optimized for the specifications of the corresponding instrument.

[0085] In some implementations, the image server system 102 can apply polygons obtained from the spatial database 108 to satellite imagery for a specific day. By analyzing the brightness or intensity of each pixel and comparing the brightness of the pixels to a threshold, the image server system 102 can determine the number of pixels that appear to indicate a wildfire within the polygons of the satellite imagery on July 5, 2019. By incorporating polygons into the satellite imagery, the image server system 102 can reduce the complexity of processing and shorten processing time because a smaller amount of pixels in the satellite imagery are searched.

[0086] The image server system 102 can calculate the brightness of each pixel in the satellite image from July 5, 2019, or in a polygon applied to the satellite image from July 5, 2019. Next, for each pixel, the image server system 102 can compare the brightness of each pixel to a threshold. For example, the image server system 102 can specify a threshold of 50 or another brightness value. If a particular pixel in the satellite image is greater than the threshold, the image server system 102 can consider that pixel to be included in the count. The image server system 102 repeats this process for each pixel in the satellite image.

[0087] In some implementations, the image server system 102 can identify pixels in satellite imagery that indicate wildfires. Depending on the context, pixels that can indicate wildfires may correspond to areas within a polygon, such as areas that are actively burning and areas of potential burning where a fire recently occurred. If an area of ​​burning where a fire recently occurred contains one or more embers from a fire, the image server system 102 can detect and identify pixels in the satellite imagery associated with those embers as pixels that are on fire. However, pixels associated with areas where no fire is occurring and scar areas resulting from a fire are not detected by the image server system 102 as indicating a fire.

[0088] Next, the image server system 102 can input the number of counted pixels in the satellite image as the value of k in equation 1 of the Poisson distribution 116. For example, the image server system 102 can fit the frequency histogram to the Poisson distribution and determine the following equation.

[0089]

number

[0090] As shown in Equation 3, the number 5 instead of the value λ corresponds to the average number of pixels that appear to indicate a wildfire from the baseline distribution for each day. Furthermore, the value λ can describe the width of the Poisson distribution curve. Equation 3 represents the probability density function or continuous probability density function of the Poisson distribution. The value k corresponds to the number of pixels found in the satellite image of July 5, 2019 that appear to indicate a wildfire. For example, if the pixel count of a particular image corresponds to a value of 6, then Equation 3 should yield the following value:

[0091]

number

[0092] As shown in Equation 4, the image server system 102 calculated a 14.62% probability that the event of six fire pixels is likely to occur in the baseline distribution. Furthermore, the calculated probability represents the number of events (e.g., fire pixels) that occurred in the satellite image in one day (e.g., July 5, 2019) relative to the known average proportion of fire pixels that occur in one year (e.g., April 5, 2018 to April 5, 2019), as determined from the baseline distribution.

[0093] In some implementations, the calculated probability can correspond to a metric that is a substitute for the probability that the satellite image is not showing a fire. Given this substitute, the image server system 102 can determine whether the satellite image shows a wildfire, depending on the threshold to which the likelihood falls. If the likelihood is greater, the likelihood indicates that the satellite image shows a wildfire. Alternatively, if the likelihood is smaller, the likelihood indicates that the satellite image does not show a wildfire.

[0094] Next, the image server system 102 can compare the output probability with a threshold to determine whether the satellite image from July 5, 2019, has a similar number of fire pixels as the baseline distribution. For example, the image server system 102 may set the threshold to 5% or 0.05. If the image server system 102 determines that the output probability is less than the threshold, it can determine that the satellite image is likely to show fire. Alternatively, if the image server system 102 determines that the output probability is greater than the threshold, it can determine that the satellite image does not contain fire.

[0095] If the image server system 102 determines that the output probability is below a threshold, it repeats the process of comparing the average number of pixels that appear to indicate a wildfire with the previous day. For example, the image server system 102 can obtain a satellite image from satellite image 114 for July 4, 2019. Then, the image server system 102 can determine the number of pixels in the satellite image for July 4, 2019 that appear to indicate a wildfire, for example, using a fire detection algorithm. The image server system 102 can determine the probability from a statistical distribution based on the count of fire pixels for that day. Then, the image server system 102 can compare the probability from the statistical distribution with a threshold to determine whether the satellite image contains a likelihood of a fire.

[0096] The image server system 102 repeats this process until the probability from the statistical distribution exceeds a threshold. In some implementations, if the image server system 102 determines that the probability of a satellite image on a particular day showing a wildfire is greater than the threshold, the image server system 102 can consider that there is no fire in this satellite image. In this case, the image server system 102 can identify a first occurrence in the satellite image that generates a probability based on the statistical distribution and the number of pixels greater than the threshold. In some implementations, the image server system 102 can consider that date to be the date the wildfire occurred. As shown in system 100, this start date 118 or the date the wildfire occurred corresponds to June 1, 2019, because the satellite image does not show an ongoing wildfire, but the satellite images for June 2, 2019, July 4, 2019, and July 5, 2019 show a fire.

[0097] In other implementations, the image server system 102 can perform additional processes if the statistical distribution outputs a probability greater than a threshold for the identified date. In particular, the image server system 102 can perform verification that the identified date on which the wildfire occurred is actually accurate and not determined by error. For example, the image server system 102 can determine that the identified date on which the wildfire occurred was not determined by noise such as clouds or smoke found in the satellite image. To verify this, the image server system 102 can calculate the probability using the statistical distribution 116 for each day one week prior to the identified start date.

[0098] For example, if the image server system 102 determines that June 1, 2019 is the identified date, the image server system 102 can perform an additional verification process to calculate probabilities using the statistical distribution 116 for each day prior to June 1, 2019, for example, going back to May 25, 2019. The image server system 102 can, for example, acquire satellite images for each day from May 25, 2019 to May 31, 2019, and count the number of fire pixels for each day from May 25, 2019 to May 31, 2019. Then, for each pixel count from each day, the image server system 102 can calculate the corresponding probability using the statistical distribution 116 from each day. If each corresponding probability is greater than a threshold, for example, if it shows a sufficiently high probability to be classified as a day without fire, the image server system 102 can consider June 1, 2019, as the true date on which the wildfire occurred. Instead, if one of the days within the previous week period shows a probability lower than a threshold, for example, if one of these days indicates a wildfire, the image server system 102 can continue iterating in reverse to find the actual start date of the wildfire and then the "no fire" days for the entire previous week. The same process applies to finding the end date of the wildfire.

[0099] In some implementations, the image server system 102 may need to make adjustments to the identified start or end date. For example, the image server system 102 may determine that the satellite cannot identify a wildfire until two days after the start date. To solve this problem, the image server system 102 can subtract two days from the identified start date to ensure accurate time range detection. For example, if the image server system 102 identifies June 1, 2019 as the start date, the image server system 102 can set the date back two days, for example, to May 30, 2019, to account for inconsistent detection with the satellite. Other days such as three, four, or five days may also be possible.

[0100] In some implementations, the image server system 102 can adjust the identified start date based on the rate of change in fire intensity. For example, if the image server system 102 notices that the satellite image has changed in intensity by 60 between one day and the next, it can move the date back by 3 days, for example, to May 29, 2019. If the image server system 102 notices that the satellite image has changed in intensity by 40 between one day and the next, it can move the date back by 4 days, for example, to May 28, 2019. Typically, rapidly spreading fires require less adjustment than slowly spreading fires. This process is similar when adjusting the identified end date.

[0101] During stage (F), the image server system 102 can generate a second statistical distribution using the wildfire start date 118. As shown from system 100, the wildfire start date 118 occurred 1 month and 4 days before the provided input date 110. Based on the second generated statistical distribution, the image server system 102 can determine the wildfire end date 124.

[0102] First, the image server system 102 can provide the satellite image database 104 with the start date 118 and location 112 in order to obtain satellite imagery for generating a second statistical distribution. Stage (F) is similar to stage (C) in that the date and location are provided to the satellite image database 104, and the satellite image database 104 returns satellite imagery indicating the location. However, in stage (F), the image server system 102 can also instruct the satellite image database 104 to provide satellite imagery from both before and after the start date 118. Based on the satellite imagery before the start date 118, the image server system 102 can generate a second statistical distribution, and based on a comparison of the satellite imagery after the start date 118 with the second statistical distribution, it can determine the end date 124 of the wildfire.

[0103] For example, the image server system 102 can provide the satellite image database 104 with a period around the start date 118 in order to acquire satellite images. The period around the start date 118 can be determined based on the amount of time required to construct a second statistical distribution, e.g., the period before the start date 118, and the amount of time required to identify the end date, e.g., the period after the start date 118. For example, the image server system 102 can identify a period of one year and three months, which is one year before the start date 118 to construct the second statistical distribution, and three months after the start date 118 to determine the end date 124 of the satellite images. In other examples, the image server system 102 can request a large amount of satellite images, such as two years, five years, or more, centered around the start date 118.

[0104] During stage (G), the satellite image database 104 can provide the identified satellite images 120a and 120b to the image server system 102 via the network. The satellite image database 104 can provide the satellite images 120a and 120b via the network via email, a zip file, or another format. In some examples, the satellite image database 104 can send an index to the image server system 102 for retrieving the satellite images 120a and 120b.

[0105] For example, satellite image 120a can correspond to satellite images acquired before the start date 118. Satellite image 120b can correspond to satellite images acquired after the start date 118. The image server system 102 can generate a second statistical distribution based on satellite image 120a. In addition, the image server system 102 can determine the end date 124 based on the generated second statistical distribution and satellite image 120b.

[0106] During stage (H), the image server system 102 can generate a second statistical distribution based on the identified satellite image 120a and the determined start date 118. Stage (H) is similar to stage (D). However, during stage (H), the image server system 102 does not need to determine a date three months prior to the input date, since the start date 118 is known. In this case, the image server system 102 can identify a time region prior to the start date 118.

[0107] In some implementations, the image server system 102 may return an error, an identified time range, or both. For example, the image server system 102 may return an error if an end date is requested for a wildfire that is still ongoing, if satellite imagery is unavailable or of poor quality for the time and geographical location of a historical fire, if satellite data is unavailable or of poor quality for the time range used to construct the noise distribution, or in other cases. The error may indicate, for example, "insufficient satellite imagery," "wildfire still ongoing," or "low-quality satellite imagery." In other examples, if the baseline distribution is too close to the current day, for example, within a day, a few days, or a week, the image server system 102 may indicate that the baseline distribution can be moved to an earlier period accordingly.

[0108] For example, the image server system 102 can generate a time range one year prior to the start date 118. In this example, the image server system 102 can generate a time range from June 1, 2018 to June 1, 2019 in order to acquire satellite images and construct a second statistical distribution. As done in step (C), the image server system 102 can ensure that there are no active wildfires in the satellite images during this time range. If the image server system 102 detects an ongoing wildfire in the satellite images, for example, or within a polygon in the satellite image, the image server system 102 can adjust the time range accordingly to avoid the satellite images showing an active wildfire.

[0109] In some implementations, the image server system 102 can generate a second statistical distribution 122 once a time range is identified. In particular, the image server system 102 can generate a second statistical distribution 122 based on a subset of satellite images 120a that match the identified time range, for example, from June 1, 2018 to June 1, 2019. The second statistical distribution 122 may also take the form of a Poisson distribution, a normal distribution, or a Gaussian distribution, to name a few examples. In some cases, the value λ for the second statistical distribution 122 may differ from the value λ for the first statistical distribution 116.

[0110] During stage (I), the image server system 102 can determine the end date 124 of the wildfire based on the generated second statistical distribution 122 and satellite imagery 120b. The function of stage (I) is the same as that of stage (E). However, the image server system 102 compares each of the satellite images from satellite imagery 120b starting from the start date of June 1, 2019, with the generated second statistical distribution 122.

[0111] For example, the image server system 102 can acquire a satellite image from June 1, 2019, from satellite image 120a. The image server system 102 can then determine the number of pixels in the June 1, 2019 satellite image that appear to indicate a wildfire by counting each of the pixels that appear to indicate a wildfire. For example, the image server system 102 can identify the number of pixels within a polygonal region of the June 1, 2019 satellite image that appear to indicate a wildfire. Based on the count of fire pixels in the image, the image server system 102 can determine a probability from a second statistical distribution 122. The image server system 102 can then determine whether the satellite image contains a likelihood of a fire by comparing the probability from the statistical distribution to a threshold.

[0112] If the image server system 102 determines that the probability of a satellite image on a particular day, for example June 2, 2019, showing a wildfire is below a threshold, the image server system 102 can consider this satellite image to appear to show a fire. In this case, the image server system 102 identifies the satellite image for the next day in the future, for example June 3, 2019, from the satellite image 120b. The image server system 102 can traverse forward in time to find the end date of the wildfire. In particular, the image server system 102 repeats this process until the probability from the statistical distribution is greater than a threshold, and the probability of any consecutive number of days immediately following the potential end date is also greater than a threshold. This number of days may be, to give a few examples, 6, 7, 8, or any other number. This process for verifying that the potential end date of the wildfire is actually correct is similar to verifying that the dates of previous potential states of the wildfire are correct. However, instead of calculating probabilities using the statistical distribution 116 for each day one week prior to the potential start date, the image server system 102 calculates probabilities using the statistical distribution 122 for each day one week after the potential end date to identify "fire-free" days in this process. The image server system 102 can continue to iterate forward to identify the actual end date of the wildfire if the next one-week period results in one or more days indicating a wildfire.

[0113] In some implementations, if the image server system 102 determines that the probability of a satellite image showing a wildfire on a particular day is greater than a threshold, the image server system 102 can consider that there is no fire in that satellite image. The image server system 102 can then repeat the process of calculating probabilities for consecutive days following the potential day, such as a week, using a statistical distribution to determine whether the potential day corresponds to the actual end date. If the image server system 102 determines that the probabilities generated for each day following the potential date are greater than a threshold, the image server system 102 can consider that date to be the date the wildfire ended.

[0114] In some implementations, the image server system 102 can determine the wildfire end date 124 by analyzing pixels within one or more burning areas and one or more scar areas resulting from the fire. The image server system 102 can determine that these corresponding pixels within the satellite image polygon correspond to either a burning area or a scar area. In addition, when the end date 124 is identified, the image server system 102 can expand the polygon area overlaid on the satellite image. By expanding, the image server system 102 can identify pixels within the maximum perimeter of the polygon to ensure that the wildfire has not spread to other areas in the satellite image.

[0115] In some implementations, the image server system 102 may include multiple surrounding polygons in satellite imagery over the lifespan of a wildfire. In this case, when constructing the statistical distribution 116, the image server system 102 can crop the polygon closest to the satellite data after the input day 110. In addition, when constructing the statistical distribution 122, the image server system 102 can crop the oldest polygon. To reduce noise when determining the start date, the image server system 102 can select the polygon closest to the statistical distribution 116 and the oldest polygon for the statistical distribution 122. If the image server system 102 is browsing satellite imagery in an area known not to have burned as a result of a wildfire during a particular day, the image server system 102 will only count noisy pixels that do not contain wildfires. Therefore, if the image server system 102 only counts fire pixels, the accuracy of the start date prediction will be reduced.

[0116] As shown in System 100, the satellite image from June 1, 2019 appears to show a wildfire, the satellite image from June 2, 2019 appears to show a wildfire, the satellite image from August 4, 2019 appears to show a smaller wildfire, and the satellite image from August 5, 2019 appears not to show a wildfire. Assuming that the image server system 102 does not detect a wildfire on any day after August 5, 2019, for example from August 6 to August 13, 2019, the image server system 102 can define the end date 124, or the date the wildfire ended, as August 5, 2019.

[0117] During stage (J), the image server system 102 can provide the output time range of the wildfire to one or more pipelines. As shown in system 100, the image server system 102 determined that the time range of the wildfire is from June 1, 2019 to August 5, 2019, based on the first statistical distribution 116, the second statistical distribution 122, and the corresponding satellite imagery. Furthermore, the image server system 102 can determine that this time range is accurate because the length of the time range is less than three months, for example, less than the maximum duration the fire lasts, and the input date 110 falls within the time range.

[0118] In some implementations, the image server system 102 can provide time ranges to one or more internal or external pipelines. Internal pipelines could, for example, be different machine learning models as training data. These different machine learning models could perform applications such as monitoring wildfires and their spread. Other internal pipelines could include, for example, different graphical representations of wildfires or other user interfaces to show the spread of wildfires. External pipelines could include, for example, other external systems that perform wildfire detection, dispatching responders to extinguish wildfires, and remotely monitoring crises such as wildfires or other climate disasters. The image server system 102 can also provide time ranges of wildfires to other systems. For example, as shown in system 100, the image server system 102 can output time ranges to monitor 126 for review.

[0119] The image server system 102 can retrieve satellite images with known start and end dates from the satellite image database 104 to verify the authenticity of the process of system 100. For example, the image server system 102 can retrieve multiple satellite images corresponding to intermediate dates within an identified time range. The image server system 102 can then provide each satellite image from the multiple satellite images to the process shown in system 100 to determine whether the image server system 102 can identify the same time range, for example, the same start and end dates. If the image server system 102 can identify the same time range for each satellite image, the process shown in system 100 appears to be functioning correctly.

[0120] Figure 2A is a block diagram showing an example of a system 200 for receiving satellite imagery and generating time ranges for wildfires from the satellite imagery. System 200 includes similar components and performs similar functions as system 100. For example, system 200 includes a satellite 202, an image server system 204, a time database 214, a spatial database 216, and a satellite imagery database 218. System 200 also includes a network 205 which can be either a satellite network, a local connection, or another connection over the internet.

[0121] During stage (A), the image server system 204 may send a request 203 to the satellite 202 via the network 205. The request 203 can instruct the satellite 202 to capture one or more satellite images of a geographic area on Earth. The request 203 may include location coordinates of the area, such as latitude and longitude coordinates, for the satellite 202 to capture one or more images as it orbits the Earth.

[0122] During stage (B), satellite 202 may capture one or more satellite images of the geographical area provided by request 203. Satellite 202 may include one or more satellite cameras for capturing satellite images. For example, as shown in system 200, the geographical area may include one or more mountains, hills, fires, and smokes. Other geographical areas may include more, fewer, or different geographical features than those shown in system 200. Satellite 202 may capture satellite images 206 of the requested geographical area as satellite 202 orbits the Earth.

[0123] For example, satellite 202 can navigate over a requested geographical area on a daily basis and therefore can provide satellite imagery of the requested geographical area on a daily basis. In some implementations, satellite 202 may capture multiple images, multiple videos, or both per day over a geographical area. In this case, the image server system 102 may be able to view multiple satellite images and videos on one day, and different sets of multiple satellite images and videos on the next day, and so on. By providing sets of media, the image server system 204 can improve the accuracy of its time-range wildfire detection.

[0124] In some implementations, one or more geographical areas captured by a satellite camera may include one or more locations associated with a wildfire. For example, one or more locations associated with a wildfire may include one or more areas where there is no fire, one or more areas that are actively burning, one or more burning areas where a fire recently occurred, and one or more scarred areas resulting from the fire. As shown in system 201, one or more locations associated with a wildfire within a geographical area may include one or more areas 211 where there is no fire, one or more areas 210 that are actively burning, one or more burning areas 208 where a fire recently occurred, and one or more scarred areas 212 resulting from the fire.

[0125] Each of regions 208, 210, 211, and 212 may be adjacent to one another within the geographic region. When satellite 202 captures images of one or more geographic regions based on the location in request 203, the satellite image may include one or more of these regions. For example, satellite 202 may capture an image including the location from request 203, and the image may include region 210 which is actively burning, burning region 208 where a fire recently occurred, and one or more scar regions 212 resulting from the fire. Other satellite images may include other regions, such as each of the regions shown in system 100.

[0126] Next, satellite 202 can transmit the captured media 206 to the image server system 204 via the network 205. In some implementations, satellite 202 can transmit the captured media 206 directly to the satellite image database 218, bypassing the image server system 204 directly. In this case, at a later point in time, the image server system 204 can obtain data identifying the captured media 206 and determine the time range of the wildfire indicated by the captured media. For example, the data identifying the captured media 206 may include, for example, location data and the specific time when the wildfire indicated by the captured media 206 was active.

[0127] During stage (C), the image server system 204 retrieves the input date 220 of the wildfire and the corresponding location 222 of the wildfire. Stage (C) of system 200 is similar to stages (A), (B), and (C) of system 100. In particular, as shown in system 200, the image server system 204 retrieves the input date 220 of July 5, 2019, and the location 222 of 33.8121N, -117.91899E for the corresponding wildfire. In some implementations, the image server system 204 can retrieve polygons to apply to the satellite image to reduce the amount of pixels to be searched. Polygons can be retrieved from the spatial database 216. Polygons can spatially limit the number of pixels searched for fire pixels by the image server system 102. For example, the image server system 204 can apply polygons 226 to the retrieved satellite image 224. The corresponding wildfire may correspond to a captured wildfire in satellite image 206, but instead, it may refer to satellite imagery previously captured and stored in satellite imagery database 218.

[0128] The image server system 204 can provide the satellite image database 218 with the date 220 on which the wildfire occurred and the location 222 on which the wildfire occurred. In addition, the image server system 204 can specify the period for which satellite image retrieval is to be performed. For example, the image server system 204 can instruct the satellite image database 218 to provide all satellite imagery starting from the date July 5, 2019, which indicates location 222, and two years prior. The image server system 204 can also request satellite imagery from the satellite image database 218 for different periods, such as one year, three years, five years, or more, to give a few examples. In response, the satellite image database 218 can return satellite imagery 224 to the image server system 204 based on the criteria specified by the image server system 204. The image server system 204 can then apply polygons 226 to the retrieved satellite imagery 224 to spatially limit the area in which the image server system 204 searches for fire pixels.

[0129] During stage (D), the image server system 204 can generate a first statistical distribution based on the acquired satellite imagery 224. Stage (D) of system 200 is the same as stage (D) of system 100.

[0130] In particular, the image server system 204 can determine a time range for identifying a subset of satellite images from the satellite image 224. The image server system 204 identifies a date three months prior to the input date 220, for example, April 5, 2019. Then, the image server system 204 determines either a one-year time range (for example, April 5, 2018 to April 5, 2019) or a nine-month time range (for example, July 5, 2019 to April 5, 2019) in order to construct a satellite image baseline.

[0131] In response to the identification of a time range, the image server system 204 can use the identified time range to identify a subset of satellite images from satellite image 224. For example, the image server system 204 can identify satellite images from satellite image 224 from April 5, 2018 to April 5, 2019, and ensure that these identified satellite images do not show wildfire pixels, for example, they do not show satellite images within a threshold. The image server system 204 can then determine the average number of fire pixels found in the identified satellite images within the time range and generate a first statistical distribution by fitting the average number of fire pixels to a statistical distribution. The statistical distribution may be, for example, a Poisson distribution where the mean is equivalent to the variance.

[0132] During stage (E), the image server system 204 can determine the number of pixels in the satellite image 224 that appear to indicate an active wildfire. Stage (E) of system 200 is similar to stage (E) of system 100. In particular, the image server system 204 can determine the number of pixels in a polygon in the satellite image that appear to indicate a wildfire on a particular day, for example, July 5, 2019, by counting the detected fire pixels, using, to give a few examples, a fire detection algorithm or using brightness threshold detection.

[0133] During stage (F), the image server system 204 can compare the number of detected fire pixels with the generated statistical distribution. Stage (F) of system 200 is the same as stage (E) of system 100. For example, if the generated statistical distribution is a Poisson distribution, the image server system 204 can determine the probability by providing the number of detected fire pixels for a particular image as a value in the Poisson distribution. The Poisson distribution can return the probability of an event occurring, for example, the probability that a particular image from July 5, 2019, shows a wildfire. The image server system 204 can compare the probability with a threshold to determine whether satellite images for a particular day have a similar number of fire pixels.

[0134] If the image server system 204 determines that the probability output by the generated statistical distribution is below a threshold, during stage (G), the image server system 204 acquires satellite imagery from the previous day. Stage (G) of system 200 is the same as stage (E) of system 100. For example, the image server system 204 acquires satellite imagery from July 4, 2019, and repeats the processes of stages (E) and (F) of system 200. The image server system 204 repeats the processes of stages (E), (F), and (G) until the start date of the wildfire is identified. Thus, the image server system 204 continues to go back one day at a time until the corresponding satellite imagery showing location 222 no longer shows an active or ongoing wildfire, for example, until the probability output by the generated statistical distribution is greater than a threshold.

[0135] Alternatively, if the image server system 204 determines that the probability output by the generated statistical distribution is greater than a threshold, during stage (H), the image server system 204 indicates that the wildfire start date 228 has been identified, assuming that the image server system 204 does not detect a wildfire within a predetermined number of days prior to the potential start date 228. As shown in system 200, the wildfire start date 228 corresponds to June 1, 2019. Stage (H) of system 200 is similar to stages (F) and (G) of system 100.

[0136] For example, during stage (H), the image server system 204 may provide the identified start date 220 and location 222 to the satellite image database 218 in order to acquire satellite imagery for generating a second statistical distribution. Here, the image server system 204 may also instruct the satellite image database 218 to provide satellite imagery from both before and after the start date 228. For example, the image server system 204 may specify a particular time frame around the start date 228 (e.g., 1 year and 3 months) or request a larger time frame around the start date 228 (e.g., 2 years, 5 years, or more).

[0137] In some implementations, the satellite image database 218 can return acquired satellite images 232 based on criteria identified by the image server system 204. In particular, the satellite images 232 may include, for example, satellite images prior to an identified start date 228 for constructing a second statistical distribution, and satellite images after an identified start date 228 for identifying, for example, the end date of a wildfire.

[0138] During stage (I), the image server system 204 can generate a second statistical distribution based on satellite imagery 232 prior to the identified start date 228 and the identified start date 228. Stage (I) of system 200 is similar to stage (H) of system 100. In this stage, the image server system 204 can identify a time range prior to the start date 228 in order to construct the second statistical distribution. Based on the identified time range, the image server system 204 can extract a subset of satellite imagery from satellite imagery 232. The image server system 204 can then fit the subset of satellite imagery to a statistical distribution such as a Poisson distribution, a normal distribution, or a Gaussian distribution.

[0139] During stage (J), the image server system 204 can determine the number of pixels in satellite imagery that appear to indicate wildfires from satellite imagery 232 after the start date 228. Stage (J) of system 200 is the same as stage (I) of system 100. In this case, the image server system 204 can count the number of pixels in satellite imagery that appear to indicate wildfires for a specific day, for example, June 1, 2019, by determining the brightness of each pixel or by using a fire detection algorithm.

[0140] During stage (K), the image server system 204 can compare the count of fire pixels for a specific image on a specific day with a second statistical distribution that has been generated. Stage (K) of system 200 is similar to stage (F) of system 200. The second statistical distribution that has been generated can generate probabilities of the likelihood of an event occurring, for example, the probability that a specific image on June 1, 2019, shows a wildfire. The image server system 204 can compare the probability with a threshold to determine whether satellite images on a particular day have a similar number of fire pixels.

[0141] If the image server system 204 determines that the probability output by the generated second statistical distribution is below a threshold, during stage (L), the image server system 204 acquires satellite imagery for the following day. Stage (L) of system 200 is the same as stage (I) of system 100. For example, the image server system 204 acquires satellite imagery for June 2, 2019, and repeats the processes of stages (J) and (K) of system 200. The image server system 204 repeats the processes of stages (J), (K), and (L), advancing by one day at a time until the end date of the wildfire is identified and the image server system 204 assumes that it has not detected a wildfire for a predetermined number of days after the potential end date.

[0142] Alternatively, if the image server system 204 determines that the probability output by the generated statistical distribution is greater than a threshold, during stage (K), the image server system 204 indicates that the wildfire end date 234 has been identified, assuming that the image server system 204 will not detect a wildfire for a predetermined number of days after the potential end date 234. As shown in system 200, the wildfire end date 234 corresponds to August 5, 2019. Stage (K) of system 200 is the same as stage (I) of system 100.

[0143] During stage (M), the image server system 204 provides a time range 236 of the wildfire for output. The time range 236 indicates that the wildfire range is from June 1, 2019 to August 5, 2019. Stage (M) of system 200 is the same as stage (J) of system 100. For example, the image server system 204 can output the time range 236 to one or more external or internal pipelines.

[0144] Figure 2B is a block diagram showing an example of system 201 for generating a statistical distribution when determining the time range of a wildfire. System 201 shows the processes performed during stages (D) and (H) from system 100 and stages (D) and (I) from system 200. Furthermore, system 201 and its processes may also be performed by image server system 102.

[0145] In some implementations, system 201 provides a process for generating a statistical distribution. The image server system 204 can generate a statistical distribution that is, to give a few examples, a Poisson distribution, a normal distribution, or a Gaussian distribution. The process for generating a statistical distribution generally includes, in particular, (i) identifying a first date that is three months prior to the input date; (ii) identifying a second date that is one year or nine months prior to the first date; (iii) selecting a subset of satellite images from acquired satellite images based on the time range between the first and second dates; (iv) determining the mean, standard deviation, and other statistical properties of fire pixels from the subset of satellite images; and (v) generating a statistical distribution based on the statistical properties of fire pixels from the subset of satellite images.

[0146] In some implementations, the image server system 204 can generate a statistical distribution without identifying a first date which is three months prior to the input date. In this case, if the image server system 204 determines that the input date corresponds to the actual start date of the wildfire, such as the start date 118 generated in system 100, the image server system 204 can proceed to identify a second date which is one year prior to the input date without identifying a first date. For example, if the input date is the determined start date of the wildfire, e.g., June 1, 2019, instead of identifying a date three months prior to June 1, 2019 to ensure that there is no wildfire in the satellite imagery, the image server system 204 can identify a second date which is one year prior to the start date. This corresponds to a second date which is June 1, 2018.

[0147] As a result, for example, if the image server system 204 sets the start date back by only three months, the image server system 204 can identify a time range of June 1, 2018 to June 1, 2019, instead of March 1, 2018 to March 1, 2019. In some cases, the image server system 204 can identify a nine-month time range instead of a one-year time range, for example, September 1, 2019 to June 1, 2019. The image server system 204 can identify a nine-month range when a shorter time range is required due to memory constraints, for example, to avoid detecting fire pixels in historical satellite imagery, or to identify a nine-month range when an external user requests that particular time range.

[0148] During stage (A), the image server system 204 can determine a first date 242 that is three months prior to the input date. For example, if the input date corresponds to the input date 110 of the wildfire, for example July 5, 2019, the image server system 204 can determine that the first date 242 is April 5, 2019. In some cases, the image server system 204 skips this stage if the input date corresponds to the start date of the wildfire, for example start date 118.

[0149] During stage (B), the image server system 204 can determine a second date 244 that is one year earlier than the first date 242. For example, if the first date 242 corresponds to April 5, 2018, and the input date corresponds to the input date 110 for the wildfire, the image server system 204 can determine that the second date 244 is April 5, 2019. However, in this example, if the input date is the start date 118, the image server system 204 may determine that the first date 242 is June 1, 2018.

[0150] During stage (C), the image server system 204 can select a subset of satellite images from the acquired satellite images. As shown in system 201, the image server system 204 can obtain location 240 of the geographical area where the wildfire occurred and date 238 of the wildfire at location 240. The date can be a string indicating, for example, July 5, 2019, and the location can be a string or number indicating the latitude and longitude coordinates 33.8121N, -117.91899E. This stage is the same as stages (A) and (B) from system 100.

[0151] Next, the image server system 204 can request satellite images from the satellite image database 218 by providing the date 238, location 240, and period 237. The period 237 can, for example, indicate a period prior to the date 238 in order to obtain satellite images. For example, the period 237 can indicate a period of one year, three years, five years, or any other time length sufficient to construct a statistical distribution.

[0152] In some implementations, the image server system 204 can indicate that period 237 corresponds to a period before date 228 and a period after date 238. For example, if the image server system 204 is trying to determine the end date of a wildfire, it can instruct the satellite image database 218 to identify two years before date 238 and five months after date 238. Other periods may also be indicated.

[0153] The satellite image database 218 may include satellite imagery depicting various geographical regions of the Earth. For example, the satellite imagery may include high-resolution images 213b, noisy images 213a, and other types of images such as infrared video and other media.

[0154] In some implementations, the image server system 204 can acquire satellite imagery and select a subset of satellite imagery from the acquired imagery. For example, as shown in system 201, the image server system 204 can select satellite imagery 213c from the acquired imagery. Satellite imagery 213c may include a set of satellite imagery showing high-resolution images, noisy-resolution images, low-resolution images, satellite video, and other satellite media. The image server system 204 can select satellite imagery 213c based on the period identified in stages (A) and (B) of system 201, one or more time ranges to avoid fire pixel detection in the set of satellite imagery, and other criteria.

[0155] During stage (D), the image server system 204 can determine the detection of fire pixels on each day over the identified time range. This stage is similar to stage (D) from system 100. For example, the image server system 204 can analyze each satellite image from satellite image 213c and determine the number of pixels that appear to indicate a wildfire on each day within the set of satellite images. For example, the image server system 102 can analyze satellite image 113c on each day within the identified time range from April 5, 2018 to April 5, 2019 and determine the number of fire pixels on each day. In another example, the image server system 102 can analyze the polygons within satellite image 113c for each day from April 5, 2018 to April 5, 2019 and determine the number of fire pixels within the polygons on each day. The image server system 102 can, for example, count two fire pixels on April 5, 2018, three fire pixels on April 6, and continue counting satellite images each day until April 5, 2019.

[0156] During stage (E), the image server system 204 can generate a statistical distribution in response to determining the number of fire pixels in the satellite image 113c for each day. For example, the image server system 204 can generate one or more parameters relating to a statistical distribution that may be a Poisson distribution, such as the mean and variance of the distribution. If the statistical distribution is a Poisson distribution, the image server system 204 can determine the average number of pixels that appear to indicate a fire from the set of satellite image 113c. In this case, the average number of identified pixels that appear to indicate a wildfire over an identified time range may be, for example, 5, and this value is equivalent to the variance of the Poisson distribution. In response to identifying the average of fire pixel detections in the satellite image and fitting that average, the image server system 204 can provide a statistical distribution 116 for various applications.

[0157] Figure 3 is a flowchart showing an example of process 300 for generating a time range for a wildfire. Image server system 102 of system 100 and image server system 204 of system 200 can perform process 300.

[0158] The image server system obtains the date on which a fire occurred within a geographical area (302). For example, an external database, such as a time database, can provide the image server system with the date on which a wildfire occurred within a geographical area. The date can be provided in month, day, year format, or another format. The date can correspond to the start date of the wildfire, the end date of the wildfire, or another date between the start and end dates of the wildfire.

[0159] In some implementations, the image server system can also retrieve locations that describe the locations where wildfires occurred, based on the date. For example, a location may include GPS coordinates, landmark names, city names, other geographical area names, specific addresses, or other location information.

[0160] The image server system can also acquire polygons that indicate areas in satellite imagery containing wildfires. These polygons can spatially limit the number of pixels in the satellite imagery that the image server system can search to identify fire pixels. Essentially, the image server system can reduce the processing and amount of pixels it needs to search to identify fire pixels by searching for fire pixels in the satellite imagery within the areas spatially defined by the polygons.

[0161] The image server system obtains first satellite imagery of the geographic area prior to the date on which the fire occurred within the geographic area (304). In some implementations, the image server system can provide the satellite image database with the obtained location describing the geographic area where the wildfire occurred, and the date on which the wildfire occurred. In addition, the image server system can also provide a time range indicating a period prior to the input date for obtaining satellite imagery. The period may include, for example, one year prior to the input date, two years prior to the input date, five years prior to the input date, or any other period prior to the input date that is sufficient to construct a statistical distribution.

[0162] A satellite image database can identify satellite imagery that meets provided criteria using the date, location, and time range of acquisition. Satellite imagery can include one or more satellite images, such as high-quality, medium-quality, low-quality, noisy, and other satellite images. Satellite imagery can represent one or more geographical areas of the Earth. In some implementations, a satellite image database can provide identified satellite imagery that meets criteria to an image server system in order to construct a statistical distribution.

[0163] The image server system selects a first set of images from first satellite imagery prior to the date on which the fire occurred within the geographical area (306). In some implementations, the image server system can identify a first set of images from first satellite imagery in order to construct a statistical distribution. First, the image server system can identify periods in the acquired satellite imagery prior to the input date that do not show wildfires. The image server system can determine a date three months prior to the input date, since wildfires generally do not last longer than three months. By identifying a date three months prior to the input date, the image server system can safely assume that the same wildfire identified by the input date and location is inactive or not burning in the corresponding satellite imagery.

[0164] In some implementations, the image server system can determine the date one year prior to a date three months prior to the input date. For example, if the input date corresponds to July 5, 2019, the image server system can determine that the date three months prior to the input date is April 5, 2019. Next, the image server system identifies the date one year prior to April 5, 2019, which corresponds to April 5, 2018. In some implementations, the image server system can use a date nine months prior instead of one year prior.

[0165] In some implementations, the image server system can use this time range to identify a first set of satellite images from a first set of satellite images in order to construct a statistical distribution. For example, the image server system can identify satellite images from satellite images obtained from a satellite image database between April 5, 2018 and April 5, 2019, including both endpoints.

[0166] However, by using the first set of satellite imagery from April 5, 2018 to April 5, 2019, the image server system can ensure that no historical fires exist in the identified satellite imagery during that period. For example, the image server system can make this determination using an external database of historical fires. If the external database of historical fires indicates that no fires existed during this time, the image server system can run a fire detection algorithm on each satellite image within the time range of April 5, 2018 to April 5, 2019 to determine the number of pixels containing fire in each image. If the number of fire pixels in the satellite imagery is below a threshold, the image server system can indicate that a particular satellite imagery does not contain fire. Alternatively, if the image server system finds that an image contains fire, for example, if the number of fire pixels is greater than a threshold, it can adjust the time range.

[0167] The image server system generates a first statistical distribution from a first set of images (308). To generate the first statistical distribution, the image server system can generate one or more parameters for the statistical distribution. For example, the image server system can determine the number of pixels that appear to indicate a fire in each day of the satellite imagery identified from (306). The image server system can use a fire detection algorithm to count the number of fire pixels in each day. In some implementations, the image server system may crop polygons for each image in the identified satellite imagery before counting fire pixels in order to reduce the number of pixels that need to be analyzed. The image server system can then determine the mean or average number of fire pixels from each image. The image server system may also determine other data descriptors, such as standard deviation and variance, to fit a particular statistical distribution.

[0168] For example, an image server system can fit the average number of fire pixels to a Poisson distribution. In a Poisson distribution, the mean is equivalent to the variance. For other distributions, such as normal and Gaussian distributions, the image server system needs to perform a different process to determine the data descriptors that help fit the identified satellite image to the other distribution. For example, once the data is fitted to a Poisson distribution, the image server system can process the acquired satellite image to identify the start date of a wildfire.

[0169] The image server system determines the start date of the fire based on a comparison of the first satellite image with the first statistical distribution (310). In some implementations, the generated statistical distribution can be used to indicate the probability that some event occurred from the baseline of the generated statistical distribution. For example, the probability density function of a Poisson distribution can return the probability that a particular day with k counted pixels occurred naturally from the noise distribution. A lower value output by the PDF indicates a higher probability that an external force, such as a wildfire, contributed to the k count. Alternatively, a higher value output by the PDF indicates a lower probability that an external force contributed to the k count, and that the image is more closely similar to an image from the baseline distribution.

[0170] The image server system can identify satellite imagery from acquired satellite imagery that dates prior to the date a wildfire occurred. The image server system can count or determine the number of pixels in the satellite imagery from the previous day that appear to indicate a fire. In some cases, the number of pixels counted may be within a polygonal region cropped by the image server system. The image server system can then provide the number of fire pixels determined for that particular day and image as input to a generated statistical distribution.

[0171] The image server system can compare the output probability from a generated statistical distribution, such as a Poisson distribution, to a threshold. If the output probability is less than the threshold, the image server system can determine that the corresponding satellite image contains a likelihood of a fire. Alternatively, if the output probability is greater than the threshold, the image server system can determine that the corresponding satellite image does not contain a fire. The image server system repeats this process if a fire is detected, going back in time, for example from July 4, 2019 to July 3, 2019, until it finds satellite images that do not appear to indicate a wildfire.

[0172] In response to the image server system determining that satellite imagery does not show a wildfire, the image server system retrieves additional satellite imagery from the acquired imagery for a predetermined number of days prior to a specific date. For example, if the image server system determines that June 1, 2019 is a potential start date, it can identify eight days prior to June 1, 2019, for example, May 24, 2019, and retrieve satellite imagery for each of those days. The image server system attempts to identify and verify whether June 1, 2019 is the actual start date, and analyzes a predetermined number of days prior to that potential start date to determine that there is no fire in the satellite imagery corresponding to those days. If there is no fire on those days, the image server system can confidently assert that June 1, 2019 is the actual start date of the fire.

[0173] First, the image server system can count the number of pixels that appear to indicate a fire on each day for a predetermined number of days prior to June 1, 2019. Next, the image server system can use a statistical distribution and the number of pixels that appear to indicate a wildfire on each corresponding day to generate a likelihood for each day. The statistical distribution can return a probability or likelihood for each day. For example, the probability might be 0.56 for May 31, 2019, 0.6 for May 30, 2019, and 0.7 for May 29, 2019. This process is repeated until May 24, 2019, or until each day for the predetermined number of days has a relevant probability. If the probability for each day is greater than a threshold, the image server system can confidently indicate that June 1, 2019, was the actual start date of the fire. If one or more of the day probabilities are below a threshold, the image server system can continue iterating in reverse to find the actual start date of the wildfire, and then find the entire week preceding the "no fire" days.

[0174] The image server system acquires second satellite imagery of the geographic area before and after the determined start date on which a fire occurred within the geographic area (312). The image server system can provide the acquired locations and identified start dates to the satellite image database. In addition, the image server system can also provide periods around the identified start date to construct a second statistical distribution, for example, a period of one year and three months, which is one year before the start date to construct the second statistical distribution and three months after the start date of the satellite imagery to determine the end date. The satellite image database can use the criteria to identify and return second satellite imagery that satisfies this criterion. (312) includes similar functionality to (304).

[0175] The image server system selects a second set of images from a second set of satellite imagery prior to the start date of the fire within the geographical area, and the second set of images is smaller than the first set of images (314). In some implementations, the image server system can identify a second set of images from the second set of satellite imagery to construct a second statistical distribution. The image server system can identify periods in the acquired satellite imagery prior to the start date that do not indicate wildfires. Since the image server system already knows the date the fire occurred, it does not need to determine a date three months prior to the start date. Therefore, the period prior to the start date for the second set of satellite imagery may be shorter than or less than the period identified for the first set of satellite imagery that included the three-month period. Alternatively, the image server system can identify a date one year prior to the start date to construct a second statistical distribution.

[0176] In some implementations, the image server system can use this time range to identify a second set of satellite images from a second set of satellite images in order to construct a second statistical distribution. For example, the image server system can identify satellite images from satellite images obtained from a satellite image database between June 1, 2018 and June 1, 2019, including both endpoints. In addition, the image server system can ensure that wildfires are not active in the second set of satellite images during the time range of June 1, 2018 to June 1, 2019. If wildfires are active, the image server system can adjust the time range. (314) includes a function similar to (306).

[0177] The image server system generates a second statistical distribution from a second set of images prior to the start date of the fire within the geographic area (316). (316) includes similar functionality to (308). In particular, the image server system can generate one or more parameters for the second statistical distribution. For example, the image server system can determine the number of pixels that appear to indicate a fire on each day of the satellite imagery identified from (314). The image server counts the number of fire pixels on each day using a fire detection algorithm and determines the average number of fire pixels from each image. The image server system may also determine other data descriptors, such as standard deviation and variance, to fit a particular statistical distribution, such as a Poisson distribution. Once the data is fitted to a second statistical distribution, such as a Poisson, the image server system can process the acquired satellite imagery to identify the end date of the wildfire.

[0178] The image server system determines the end date of the fire based on a comparison of a second satellite image acquired after the start date with a second statistical distribution (318). Paragraph (318) includes a function similar to that of (310). In some implementations, the image server system can identify satellite images from satellite images acquired on days after the start date. The image server can count the number of fire pixels in the identified satellite image and provide the determined number of fire pixels as input to the second statistical distribution. If the probability output from the second statistical distribution is less than a threshold, the image server system can determine that the corresponding satellite image contains the likelihood of a fire. Alternatively, if the output probability is greater than the threshold, the image server system can determine that the corresponding satellite image does not contain a fire. The image server system repeats this process, going forward one day at a time, for example, from June 1, 2019 to June 2, 2019 to June 3, 2019, until it finds a satellite that does not appear to indicate a wildfire.

[0179] In response to detecting a date as a potential end date, for example, August 5, 2019, the image server system can acquire additional satellite imagery from the acquired satellite imagery for a predetermined number of days after the potential end date. Here, the image server system attempts to identify and verify whether August 5, 2019 is the actual end date, and analyzes the predetermined number of days after the potential end date to determine that there is no fire in the satellite imagery corresponding to those days. If there is no fire on those days, the image server system can confidently assert that August 5, 2019 is the actual end date of the wildfire.

[0180] Similar to (310), the image server system can count the number of pixels that appear to indicate a wildfire on each day for a given number of days after August 5, 2019. The image server system can use a second statistical distribution and the number of pixels that appear to indicate a wildfire on each corresponding day to generate a likelihood for each day. The statistical distribution can return a probability or likelihood for each day. If the probability for each day is greater than a threshold, the image server system can confidently indicate that August 5, 2019 was the actual end date of the fire. If one or more of the probabilities for the days are less than the threshold, the image server system can continue to iterate forward in time to find the actual end date of the wildfire and then find the entire following week of "no fire" days.

[0181] The image server system provides a time range for output, including the start and end dates of fires within a geographical area (320). In some examples, the time range of a wildfire, e.g., the determined start and end dates, can be provided to one or more other pipelines and can also be provided to the display of the image server system. In addition, the image server system can provide the time range to the client devices of users who have requested it via the network.

[0182] All embodiments and functional operations of the present invention described herein may be implemented in digital electronic circuits, or in computer software, firmware, or hardware including the structures disclosed herein and their structural equivalents, or in combination of one or more thereof. Embodiments of the present invention may be implemented as one or more computer program products, i.e., one or more modules of computer program instructions encoded on a computer-readable medium for execution by a data processing device or for controlling the operation of a data processing device. The computer-readable medium may be a non-temporary computer-readable storage medium, a machine-readable storage device, a machine-readable storage substrate, a memory device, a composition of a material that produces a machine-readable propagating signal, or one or more of these. The term “data processing device” encompasses all devices and machines for processing data, including, for example, a programmable processor, a computer, or multiple processors or computers. In addition to hardware, a device may include code that makes up the execution environment for the computer program in question, such as processor firmware, a protocol stack, a database management system, an operating system, or one or more of these. A propagated signal is an artificially generated signal, such as a mechanically generated electrical signal, optical signal, or electromagnetic signal generated to encode information for transmission to a suitable receiver device.

[0183] Computer programs (also known as programs, software, software applications, scripts, or code) may be written in any form of programming language, including compiled or interpreted languages, and may be deployed as standalone programs or in any form, including modules, components, subroutines, or other units suitable for use in a computing environment. Computer programs do not necessarily correspond to files in a file system. A program may be stored in part of a file that holds other programs or data (e.g., one or more scripts stored in a markup language document), in a single file dedicated to the program in question, or in a set of coordinated files (e.g., a file storing one or more modules, subprograms, or parts of code). Computer programs 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 communication network.

[0184] The processes and logic flows described herein may be implemented by one or more programmable processors that execute one or more computer programs to perform their functions by acting on input data and producing outputs. The processes and logic flows may also be implemented by dedicated logic circuits, such as FPGAs (field programmable gate arrays) or ASICs (application-specific integrated circuits), and the devices may also be implemented as dedicated logic circuits, such as FPGAs or ASICs.

[0185] Processors suitable for executing computer programs include, for example, both general-purpose and dedicated microprocessors, as well as any one or more processors in any type of digital computer. Generally, a processor receives instructions and data from read-only memory, random-access memory, or both. Essential elements of a computer are a processor for executing instructions, and one or more memory devices for storing instructions and data. Generally, a computer also includes one or more mass storage devices for storing data, such as magnetic disks, magneto-optical disks, or optical disks, or is operable to receive data from or transfer data to or from one or more mass storage devices, or both. However, a computer is not required to have such devices. Furthermore, a computer can be incorporated into other devices, for example, a few examples including tablet computers, mobile phones, personal digital assistants (PDAs), mobile audio players, and Global Positioning System (GPS) receivers. Computer-readable media suitable for storing computer program instructions and data include, for example, 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. Processors and memory may be complemented by or incorporated into dedicated logic circuits.

[0186] To provide user interaction, embodiments of the present invention may be implemented on a computer, which has 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 (e.g., mouse, trackball) on which the user can provide input to the computer. User interaction may be provided using other types of devices, 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 the input from the user may be received in any form, including acoustic, voice, or tactile input.

[0187] Embodiments of the present invention may be implemented in a computing system including a backend component (e.g., as a data server), a middleware component (e.g., an application server), or a frontend component (e.g., a client computer having a graphical user interface or web browser on which a user can interact with an implementation of the present invention), or in any combination of one or more such backend, middleware, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication, such as a communication network. Embodiments of a communication network include a local area network ("LAN") and a wide area network ("WAN"), such as the Internet.

[0188] A computing system may include clients and servers. Clients and servers are generally geographically separated from each other 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.

[0189] While several implementations have been described in detail above, other modifications are possible. For example, although the client application is described as accessing the delegate, in other implementations, the delegate may be employed by other applications implemented by one or more processors, such as applications running on one or more servers. In addition, the logic flow depicted in the diagram does not require a specific order or sequence shown to achieve the desired result. Furthermore, other actions may be provided from the described flow, or actions may be removed from the described flow, and other components may be added to or removed from the described system. Thus, other implementations are within the scope of the following claims.

[0190] This specification includes details of many specific implementations, but these should not be interpreted as limiting the scope of any invention or claimed subject matter, but rather as descriptions of features that may be specific to a particular embodiment of a particular invention. 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 also 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 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.

[0191] 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 modules and components in the embodiments described above should not be understood as requiring such separation in all embodiments, and the described program components and systems can generally be integrated together in a single software product or packaged in multiple software products.

[0192] Specific embodiments of this subject matter have been described. Other embodiments are within the scope of the following claims. For example, the actions enumerated in the claims can be performed in a different order and still achieve the desired results. As an example, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain implementations, multitasking and parallel processing may be advantageous.

Claims

1. A computer implementation method, To obtain a first satellite image of the geographic area prior to the date on which the fire occurred within the geographic area, Selecting a first set of images from the first satellite image, wherein the first set of images is selected such that i) it was captured before the date on which the fire occurred in the geographic area, and ii) it does not show any wildfires occurring in the geographic area. To generate a first statistical distribution from the first set of images, The start date of the fire is determined using the first satellite image and the first statistical distribution. i) to acquire second satellite images of the geographical area captured before the determined start date and ii) after the determined start date, Select a second set of images from the second satellite image prior to the determined start date, To generate a second statistical distribution from a second set of images prior to the determined start date, The end date of the fire is determined using the second satellite image and the second statistical distribution. A computer implementation method comprising providing, for output, the start date and end date of the fire within the geographical area.

2. The computer implementation method according to claim 1, wherein generating the first statistical distribution includes generating a first model, and generating the second statistical distribution includes generating a second model.

3. The computer implementation method according to claim 2, wherein the first model and the second model are frequency models of satellite noise in a specific geographical area over a certain period of time.

4. The computer implementation method according to claim 2 or 3, wherein the statistical distribution is modeled using a Poisson distribution, a Gaussian distribution, or a normal distribution.

5. To obtain the date on which the fire occurred within the aforementioned geographical area, Obtaining a location that describes the aforementioned geographical area, Obtaining the date on which the fire occurred within the geographical area, wherein the date may correspond to the start date of the fire, the end date of the fire, or another date within the range, To obtain the first satellite image of the geographic area prior to the date on which the fire occurred within that geographic area, To provide the location describing the geographical area where the fire occurred, and the date on which the fire occurred, to a satellite image database, A computer implementation method according to any one of claims 1 to 4, comprising: in response to providing the location and the date, obtaining the first satellite image showing the geographic area from a period prior to the date.

6. Selecting a first set of images from the first satellite images prior to the date on which the fire occurred within the geographical area, Identifying a period for generating the first statistical distribution, wherein the period corresponds to a period prior to the date, Based on the aforementioned period, a first set of images is selected from the first satellite imagery, Determining whether one or more pixels from the first set of selected images indicate a fire, A computer implementation method according to any one of claims 1 to 5, comprising adjusting a first set of selected images by comparing one or more pixels indicating a fire with a threshold.

7. To generate the first statistical distribution from the first set of images, In response to adjusting the first set of selected images, one or more parameters of the first statistical distribution are set. Determine the number of pixels indicating a fire in each day of the first set of the aforementioned images, Determining the average number of pixels indicating a fire on each day in the first set of the aforementioned images, The computer implementation method according to claim 4, comprising generating the first statistical distribution by setting the determined mean to one or more parameters.

8. Based on the comparison between the first satellite image and the first statistical distribution, the start date of the fire can be determined. Identifying satellite images from the first satellite image corresponding to a day prior to the date on which the fire occurred, From the identified satellite image from the aforementioned day, determine the number of pixels that appear to indicate a fire, Based on the first statistical distribution and the number of pixels determined to appear to indicate a fire, the likelihood that the satellite image from the first satellite image includes an indication of a fire is generated. Comparing the aforementioned likelihood with a threshold, A computer implementation method according to any one of claims 1 to 7, comprising: in response to determining that the likelihood does not exceed the threshold, obtaining another satellite image from the first satellite image corresponding to another day prior to the date on which the fire occurred, in order to determine the start date.

9. In response to determining that the likelihood for a particular day exceeds the threshold, the method Acquiring additional satellite images for a predetermined number of days prior to the aforementioned specific day, For each of the aforementioned predetermined number of days, To determine the number of active fire pixels from the aforementioned additional satellite images for that day, Based on the first statistical distribution and the number of pixels determined to appear to indicate a fire, a likelihood is generated that the additional satellite imagery for that day does not indicate a fire. The computer implementation method according to claim 8, further comprising: determining, in response to determining that each of the predetermined number of days does not indicate the fire, that the day corresponding to the identified satellite image corresponds to the start date of the fire.

10. To acquire the second satellite image of the geographic area before the start date on which the fire was determined to have occurred and after the start date on which the fire was determined to have occurred within the geographic area, To provide the location describing the geographical area where the fire occurred, and the start date of the fire, to the satellite image database, A computer implementation method according to any one of claims 1 to 9, comprising: in response to providing the location and the start date, obtaining the second satellite image showing the geographic area from periods before and after the start date.

11. Selecting a second set of images from the second satellite imagery prior to the start date on which the fire occurred within the geographical area, wherein the amount of the second set of images is less than the amount of the first set of images, Identifying a period for generating the second statistical distribution, wherein the period corresponds to a period prior to the start date. Based on the aforementioned period, a second set of images is selected from the second satellite imagery, Determining whether one or more pixels from the second set of selected images indicate a fire, A computer implementation method according to any one of claims 1 to 10, comprising adjusting a second set of selected images by comparing one or more pixels indicating a fire with a threshold.

12. From a second set of images prior to the start date on which the fire occurred within the geographical area, the second statistical distribution is generated. In response to adjusting the second set of selected images, one or more parameters of the second statistical distribution are set. Determine the number of pixels indicating a fire in each day of the second set of the aforementioned images, Determining the average number of pixels indicating a fire on each day in the second set of the aforementioned images, The computer implementation method according to claim 11, comprising generating the second statistical distribution by setting the determined mean to one or more parameters.

13. The end date of the fire can be determined based on a comparison of the second satellite image and the second statistical distribution that occurred after the aforementioned start date. Identifying satellite images from the second satellite image corresponding to a date after the aforementioned start date, From the identified satellite image from the aforementioned day, determine the number of pixels that appear to indicate a fire, Based on the second statistical distribution and the number of pixels determined to appear to indicate a fire, the likelihood that the satellite image from the second satellite image includes an indication of a fire is generated. Comparing the aforementioned likelihood with a threshold, In response to determining that the likelihood exceeds the threshold, the process includes obtaining another satellite image from the first satellite image corresponding to a different day after the date on which the fire occurred, in order to determine the end date. In response to determining that the likelihood for a particular day exceeds the threshold, the method Acquiring additional satellite images for a predetermined number of days following the aforementioned specific day, For each of the aforementioned predetermined number of days, To determine the number of pixels from the aforementioned additional satellite images for that day, Based on the second statistical distribution and the number of pixels determined to appear to indicate a fire, a likelihood is generated that the additional satellite imagery for that day does not indicate a fire. The computer implementation method according to any one of claims 1 to 12, further optionally comprising: determining that each of the predetermined number of days does not indicate the fire, and determining that the day corresponding to the identified satellite image corresponds to the end date of the fire.

14. It is a system, A system comprising one or more computers and one or more storage devices for storing instructions, wherein when an instruction is executed by the one or more computers, the one or more computers are operable to cause the one or more computers to carry out the method according to any one of claims 1 to 13.

15. Software, optionally stored on a non-temporary computer-readable medium, which includes instructions executable by one or more computers, wherein the instructions, at the time of execution, cause the one or more computers to perform the method according to any one of claims 1 to 13.

Images