Mountain fire spreading prediction method and device, equipment and storage medium

Abstract

The invention discloses a forest fire spreading prediction method, device and equipment and a storage medium, and belongs to the technical field of forest fire monitoring, and the method comprises the steps: obtaining an environment feature set of a first region based on a remote sensing image of the first region, the environment feature set comprising elevation data, gradient, slope direction, vegetation type, combustible water content, wind speed, wind direction angle, temperature and relative humidity; based on the environment feature set, a fire source feature vector of the forest fire in the first area is obtained, and the fire source feature vector comprises first fire intensity at the fire breakout moment; on the basis of the environment feature vector and the fire source feature vector, a forest fire spreading state vector is constructed, and the forest fire spreading state vector comprises a live line coordinate set, a live line spreading speed and a live line spreading direction; and based on the live wire spreading speed and the live wire spreading direction, updating a live wire coordinate set in the forest fire spreading state vector so as to predict the forest fire in the first area. The method can accurately predict the trend of forest fire spreading.

Description

Technical Field

[0001] The present disclosure relates to the technical field of wildfire monitoring, and particularly to a wildfire spread prediction method, device, equipment and storage medium. Background Art

[0002] Wildfires are one of the natural disasters that seriously threaten the forest ecosystem and the safety of human life and property. Accurately predicting the wildfire spread trend is of great significance for fire prevention and control and emergency rescue.

[0003] Traditional wildfire spread prediction methods mainly rely on single data sources and simple models, and it is difficult to accurately reflect the interaction relationship between wildfires and the environment. Summary of the Invention

[0004] The present disclosure provides a wildfire spread prediction method, device, equipment and storage medium, which can accurately predict the wildfire spread trend. The technical solution at least includes the following solutions: In a first aspect, a wildfire spread prediction method is provided, including: obtaining an environmental feature set of a first region based on a remote sensing image of the first region, where the first region is a region where a wildfire exists, and the environmental feature set includes elevation data, slope, aspect, vegetation type, fuel moisture content, wind speed, wind direction angle, temperature, relative humidity; obtaining a fire source feature vector of the wildfire in the first region based on the environmental feature set, where the fire source feature vector includes a first fire intensity at the ignition time; constructing a wildfire spread state vector based on the environmental feature vector and the fire source feature vector, where the wildfire spread state vector includes a set of fire line coordinates, a fire line spread speed, and a fire line spread direction, and the fire line spread speed is determined based on the first fire intensity; updating the set of fire line coordinates in the wildfire spread state vector based on the fire line spread speed and the fire line spread direction to predict the wildfire in the first region.

[0005] Optionally, the wildfire spread state vector is represented by the following formula:

[0006] Wherein, is the wildfire spread state vector at time is the set of fire line coordinates at time is the fire line spread speed of each fire line coordinate in the set of fire line coordinates at time and is the fire line spread direction of each fire line coordinate in the set of fire line coordinates at time ​​

[0007] Optionally, representing any one of the fire line coordinates in the fire line coordinate set at a certain moment in the set, , the fire line coordinate at a certain moment the fire line spread speed at the location is calculated using the following formula:

[0008] the fire line coordinate at a certain moment the fire line spread direction at the location is calculated using the following formula:

[0009] wherein, is the fire line spread speed at the fire line coordinate at a certain moment at the location, is the slope at the fire line coordinate at a certain moment at the location, is the wind speed at the fire line coordinate at a certain moment at the location, is the wind direction angle at the fire line coordinate at a certain moment at the location, is the slope aspect at the fire line coordinate at a certain moment at the location, is the normalized water content index at the fire line coordinate at a certain moment at the location, is the relative humidity at the fire line coordinate at a certain moment at the location, is the water content of combustibles at the fire line coordinate at a certain moment at the location, , , , , , , are weights.

[0010] Optionally, updating the fire line coordinate set in the wildfire spread state vector includes: Updating the fire line coordinate set in the wildfire spread state vector is achieved using the following formula:

[0011] Among them, is the set of fire line coordinates in the updated wildfire spread state vector, representing the set of fire line coordinates at time is the length of the time step, representing the time difference between time and time

[0012] Optionally, the method further includes: generating a wildfire report based on the updated set of fire line coordinates, and the wildfire report further includes the real-time fire intensity of the wildfire, and the real-time fire intensity is calculated based on the set of environmental parameters.

[0013] Optionally, the method further includes: correcting the updated set of fire line coordinates and the real-time fire intensity in the wildfire report.

[0014] In a second aspect, a wildfire spread prediction device is further provided, including: a first acquisition module, configured to acquire a set of environmental characteristics of the first area based on a remote sensing image of the first area, where the first area is an area where a wildfire exists, and the set of environmental characteristics includes elevation data, slope, aspect, vegetation type, water content of combustibles, wind speed, wind direction angle, temperature, relative humidity; a second acquisition module, configured to acquire a fire source feature vector of the wildfire in the first area based on the set of environmental characteristics, where the fire source feature vector includes the first fire intensity at the ignition time; a state vector construction module, configured to construct a wildfire spread state vector based on the environmental feature vector and the fire source feature vector, where the wildfire spread state vector includes a set of fire line coordinates, a fire line spread speed, and a fire line spread direction, and the fire line spread speed is determined based on the first fire intensity; an update module, configured to update the set of fire line coordinates in the wildfire spread state vector based on the fire line spread speed and the fire line spread direction, so as to predict the wildfire in the first area.

[0015] Optionally, the update module is further configured to update the set of fire line coordinates in the wildfire spread state vector by using the following formula:

[0016] Among them, is the set of fire line coordinates in the updated wildfire spread state vector, representing the set of fire line coordinates at time is the length of the time step, representing the time difference between time and time

[0017] Optionally, the device further includes: a generation module, configured to generate a wildfire report based on the updated set of fire line coordinates, where the wildfire report further includes the real-time fire intensity of the wildfire, and the real-time fire intensity is calculated based on the set of environmental parameters.

[0018] Optionally, the device further includes: a correction module, configured to correct the updated set of fire line coordinates and the real-time fire intensity in the wildfire report.

[0019] In a third aspect, a computer device is further provided, including: a memory and a processor, where at least one computer program is stored in the memory, and the at least one computer program is loaded and executed by the processor, so as to execute the wildfire spread prediction method described in the foregoing embodiments.

[0020] In a fourth aspect, a computer-readable storage medium is further provided, where at least one computer program is stored in the computer-readable storage medium, and the at least one computer program is loaded and executed by a processor, so as to execute the wildfire spread prediction method described in the foregoing embodiments.

[0021] In a fifth aspect, a computer program product is provided, including computer programs / instructions, where when the computer programs / instructions are executed by a processor, the method described in the first aspect is implemented.

[0022] The beneficial effects brought by the technical solutions provided in the embodiments of the present disclosure at least include: In the embodiments of the present disclosure, by obtaining the set of environmental features of the first area from the remote sensing image of the first area, and then constructing a wildfire spread state vector based on the set of environmental features, where the wildfire spread state vector includes a set of fire line coordinates, a fire line spread speed, and a fire line spread direction, by using the fire line spread speed and the fire line spread direction to update the set of fire line coordinates, it is possible to accurately predict the set of fire line coordinates. And because the set of environmental features includes elevation data, slope, aspect, vegetation type, fuel moisture content, wind speed, wind direction angle, temperature, and relative humidity, it is equivalent to using multiple parameters to jointly update the set of fire line coordinates. On the premise of the joint action of multiple parameters, the updated set of fire line coordinates can also have a high accuracy in a complex environment. Description of the Drawings

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present disclosure. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0024] Figure 1The flowchart of the wildfire spread prediction method provided by an exemplary embodiment of the present disclosure is shown; Figure 2 The flowchart of the wildfire spread prediction method provided by another exemplary embodiment of the present disclosure is shown; Figure 3 The structural schematic diagram of the wildfire spread prediction device provided by an exemplary embodiment of the present disclosure is shown; Figure 4 It is the structural schematic diagram of the computer device provided by the embodiment of the present disclosure. Detailed implementation manners

[0025] Unless otherwise defined, the technical terms or scientific terms used herein shall have the ordinary meanings understood by those of ordinary skill in the art to which the present disclosure pertains. The "first", "second", "third" and similar terms used in the specification and claims of the patent application of the present disclosure do not denote any order, quantity or importance, but are only used to distinguish different components. Similarly, the terms such as "a" or "one" do not denote a quantity limitation, but mean that there is at least one. The terms such as "comprising" or "including" mean that the elements or objects appearing before "comprising" or "including" cover the elements or objects listed after "comprising" or "including" and their equivalents, and do not exclude other elements or objects.

[0026] To make the objectives, technical solutions and advantages of the present disclosure clearer, the embodiments of the present disclosure will be further described in detail below with reference to the accompanying drawings.

[0027] Figure 1 The flowchart of the wildfire spread prediction method provided by an exemplary embodiment of the present disclosure is shown. This method can be executed by a computer device. Refer to Figure 1 and this method includes: In step 101, based on the remote sensing image of the first area, an environmental feature set of the first area is obtained.

[0028] The first area is the area where a wildfire exists. The environmental feature set includes elevation data, slope, aspect, vegetation type, fuel moisture content, wind speed, wind direction angle, temperature, and relative humidity.

[0029] Here, the application scenario of the embodiment of the present disclosure is that remote sensing satellites will conduct all-round and continuous monitoring of global regions. Therefore, in the case of a wildfire breaking out in any area, the remote sensing satellite can obtain the remote sensing images of this area before the wildfire breaks out and during the wildfire combustion process, and the remote sensing satellite will also send the obtained remote sensing images of this area to the computer device in real time.

[0030] By monitoring the remote sensing images in the computer device, the staff can determine the area where the wildfire currently exists, and then set a certain area as the first area. At this time, the computer device includes multiple remote sensing images of the first area.

[0031] Based on multiple remote sensing images of the first area, a DEM (Digital Elevation Model) of the first area can be established. When dealing with the problem of a small-scale wildfire, the surface of the earth in the studied area (the first area) can be regarded as a plane. In this case, the DEM data of the first area can be represented by a projected coordinate system (PCS).

[0032] The first area in the DEM can be divided into several discrete regular grids, an appropriate grid size is defined, and a grid index system is established, where each grid point is a set of coordinate data, that is, the coordinates of the first area are discretized.

[0033] Any set of coordinate data in DEM includes: X coordinate (Easting): eastward distance; Y coordinate (Northing): northward distance; Z coordinate (Elevation): elevation value. In this DEM data, the Z coordinate is a binary function of the X coordinate and the Y coordinate. In other words, any set of DEM data can be expressed as the coordinates on the X coordinate axis. , the coordinate on the Y axis To express.

[0034] In this case, optionally, at the coordinates of the first region The set of environmental features can be expressed as the following formula (1).

[0035] (1) In formula (1), for The elevation data at , that is, the elevation value, for The slope of for The slope direction, for The vegetation type of the area. express Wind speed Wind direction A collection of for The temperature at for Relative humidity at the place; for The water content of combustibles at can be calculated using the normalized water content index at and the relative humidity .

[0036] The method for obtaining each parameter in formula (1) is described below.

[0037] Optionally, the slope at is calculated using formula (2).

[0038] (2) In formula (2), denotes the partial derivative of with respect to and denotes the partial derivative of

[0039] (3) (4) In formulas (3) and (4), denotes the step size used when calculating the partial derivative. The meanings of other parameters in formulas (3) and (4) are the same as those in formulas (1) and (2), and are not elaborated here.

[0040] Optionally, the aspect at is calculated using formula (5).

[0041] (5) In formula (5), is the two-parameter arctangent function. The meanings of other parameters in formula (5) are the same as those in formulas (1) to (4), and are not elaborated here. There are many implementation methods of the two-parameter arctangent function in related technologies, and are not elaborated here.

[0042] Optionally, the vegetation type at can be determined according to the normalized difference vegetation index at . Among them, the normalized difference vegetation index at

[0043] (6) In formula (6), is the normalized difference vegetation index at is the near-infrared band reflectance at is the red band reflectance at and can both be obtained from remote sensing images.

[0044] Vegetation type includes four types: bare land or sparse vegetation, shrubs or sparse woodlands, medium-density woodlands, and high-density woodlands. In the case of it indicates that is bare land or sparse vegetation; in the case of it indicates that is shrubs or sparse woodlands; in the case of it indicates that is medium-density woodland; in the case of it indicates that is high-density woodland.

[0045] Optionally, the water content of combustibles at can be calculated using the normalized water content index at and the relative humidity at where the normalized water content index at is calculated using formula (7),

[0046] (7) In formula (7), is the normalized water content index at is the short-wave infrared band reflectance at which can be obtained from remote sensing images. The meanings of other parameters in formula (7) are the same as those in formula (6) and are not elaborated here.

[0047] (8) In formula (8), is the weight of the relative humidity at The meanings of other parameters in formula (8) are the same as those in formula (1) and formula (7) and are not elaborated here.

[0048] In some cases, such as when there are water bodies such as rivers and lakes at a certain location, the normalized water content index

[0049] should be set to the maximum value, or it can be directly considered that it is impossible to catch fire at these locations, because the fire line will not spread to water bodies such as rivers and lakes. At the location of the wind speed and the wind direction angle in the set of at the location of the wind speed and the wind direction angle

[0050] at the location of the temperature can be calculated from the thermal infrared remote sensing data at the location of

[0051] at the location of the relative humidity

[0052] In step 102, based on the set of environmental characteristics, the fire source feature vector of the wildfire in the first region is obtained.

[0053] The fire source feature vector includes the first fire intensity at the ignition moment.

[0054] Remote sensing satellites can obtain remote sensing images of the first region before and during the wildfire. Usually, the remote sensing images contain thermal infrared remote sensing data, and the thermal infrared remote sensing data can reflect temperature information. Since the temperature difference before and after ignition is large, when the remote sensing images of the first region before and during the wildfire are both known, the ignition point (i.e., the fire source location) of the first region and the relevant parameters of the ignition point can be more accurately located, so that the first fire intensity at the ignition moment of the first region can be determined.

[0055] Optionally, when the fire source location is at the ignition moment the first fire intensity

[0056] (9) In formula (9), is the first fire intensity at the ignition moment indicating the ignition moment Temperature at the position of the lower heat source , indicating the fire starting moment Temperature at the background position of the first lower area (the average value can be taken), where the background position is the position in the first area except the fire starting point. 、 are weights.

[0057] Among them, and can both be calculated from the thermal infrared remote sensing data of the first area at the fire starting moment . For example, the formula (10) is used to calculate the temperature through the thermal infrared remote sensing data.

[0058] (10) In the formula (10), is the absolute temperature calculated from the thermal infrared remote sensing data, and the unit is K (Kelvin). 、 are two constants in Planck's radiation law, , ; among them is Planck's constant, is the speed of light in a vacuum, is Boltzmann's constant. is the observation wavelength, and the unit of the observation wavelength is m (meter), radiation intensity of the black body at the wavelength .

[0059] In this case, exemplarily, the fire source feature vector can be expressed as .

[0060] In step 103, based on the environmental feature vector and the fire source feature vector, a wildfire spread state vector is constructed.

[0061] The wildfire spread state vector includes a set of fire line coordinates, the fire line spread speed, and the fire line spread direction. The fire line spread speed is determined based on the first fire intensity.

[0062] In step 104, based on the fire line spread speed and the fire line spread direction, the set of fire line coordinates in the wildfire spread state vector is updated to predict the wildfire in the first area.

[0063] In the embodiments of the present disclosure, by obtaining the environmental feature set of the first area from the remote sensing image of the first area, and then constructing the wildfire spread state vector based on the environmental feature set, the wildfire spread state vector includes the fire line coordinate set, the fire line spread speed, and the fire line spread direction. By updating the fire line coordinate set with the fire line spread speed and the fire line spread direction, it is possible to accurately predict the fire line coordinate set. Moreover, since the environmental feature set includes elevation data, slope, aspect, vegetation type, fuel moisture content, wind speed, wind direction angle, temperature, and relative humidity, it is equivalent to using multiple parameters to jointly update the fire line coordinate set. Under the premise of the joint action of multiple parameters, the updated fire line coordinate set can also have a high accuracy in a complex environment.

[0064] Figure 2 FIG. shows a flowchart of a wildfire spread prediction method provided by another exemplary embodiment of the present disclosure. This method can be executed by a computer device. Refer to Figure 2 , this method includes: In step 201, based on the remote sensing image of the first area, obtain the environmental feature set of the first area.

[0065] In step 202, based on the environmental feature set, obtain the fire source feature vector of the wildfire in the first area. For the relevant content of steps 201 to 202, refer to the foregoing steps 101 to 102, and the details are omitted here.

[0066] In step 203, based on the environmental feature vector and the fire source feature vector, construct the wildfire spread state vector.

[0067] The wildfire spread state vector includes the fire line coordinate set, the fire line spread speed, and the fire line spread direction. The fire line spread speed is determined based on the first fire intensity.

[0068] Optionally, the wildfire spread state vector is represented by the following formula:

[0069] Wherein, is the wildfire spread state vector at time, is the fire line coordinate set at time, is the fire line spread speed of each fire line coordinate in the fire line coordinate set at time, is the fire line spread direction of each fire line coordinate in the fire line coordinate set at time.

[0070] Here The moment can be the current moment or a certain moment in history. It can be an array containing all the fire line coordinates of the fire line at the moment, which can be obtained through the remote sensing image of the first area at the moment. Here, the boundary of the fire line and the burning area is not the same concept. The fire line refers to the part of the boundary of the burning area that can continue to burn outward and affect the unburned area. There will also be a part in the boundary of the burning area that cannot continue to burn. For example, when water bodies such as rivers and lakes are located at the boundary of the burning area, the fire line does not include the part of the water bodies such as rivers and lakes.

[0071] Exemplarily, denotes any fire line coordinate in the set of fire line coordinates at the moment , that is . In this case, the fire line spread speed at the fire line coordinate in at the moment is calculated using formula (11). The fire line spread direction at the fire line coordinate in at the moment is calculated using formula (12). In formulas (11) and (12),

[0072] (11) (12) In formulas (11) and (12), is the fire line spread speed at the fire line coordinate at the moment, is the fire line spread direction at the fire line coordinate at the moment, is the fire line spread speed at the fire line coordinate at the moment, where is an integer and is greater than 0. is the slope at the fire line coordinate at the moment, which is calculated using formula (2). is the wind speed at the fire line coordinate at the moment, is the wind direction angle at the fire line coordinate at the moment. is the slope direction at the fire line coordinate at, calculated using formula (5). is the normalized water content index at the fire line coordinate at, calculated using formula (7). is the relative humidity at the fire line coordinate at, is the fuel moisture content at the fire line coordinate at, calculated using formula (8); , , , , , , are weights.

[0073] Among them, , can be obtained from the data of near-surface meteorological observation stations, or retrieved from meteorological radar data, or obtained from satellite remote sensing wind speed products such as scatterometer data. can also be obtained from meteorological observation data.

[0074] In the above formula (11), the slope affects the wildfire burning speed, which is different on uphill, downhill, and flat ground; the wind speed has an additive effect on wildfire burning, and the slope direction affects the wind direction angle. For example, in the part close to the ground, the wind direction angle is actually equal to the slope direction; the fuel moisture content has an inhibitory effect on wildfire burning (the higher the fuel moisture content, the lower the burning efficiency), and the relative humidity also has a certain inhibitory effect on wildfire burning. The wildfire spread direction is related to the slope and the wind direction angle.

[0075] Therefore, based on the environmental characteristics in formula (11) and formula (12), the wildfire spread speed and the wildfire spread direction can be calculated.

[0076] Let the ignition time be time, and the current time be time . The remote sensing satellite sends remote sensing images of the first area at each time step. Then, from time until the current time, each time step can be successively represented as , , …… .

[0077] It can be seen from formula (11) that the wildfire spread speed at time is based on The fire spread speed at a moment is calculated based on The fire spread speed at a moment is based on The fire spread speed at a moment is calculated based on... The fire spread speed at a moment is based on The fire spread speed at a moment is calculated based on. Therefore, The fire spread speed at a moment is related to The fire spread speed at a moment.

[0078] Optionally, The fire spread speed at a moment Is calculated using formula (13).

[0079] (13) In formula (13), Is The fire spread speed at the fire source position at a moment, Is the weight. The meanings of other parameters in formula (13) are the same as those in formula (9) and are not elaborated here.

[0080] Through formulas (11) to (13), it is possible to For any fire line position in the fire line coordinate set at a moment Calculate the corresponding fire spread speed and fire direction, thereby constructing a wildfire spread vector.

[0081] In step 204, based on the fire spread speed and fire spread direction, update the fire line coordinate set in the wildfire spread state vector to predict the wildfire in the first area.

[0082] Optionally, use formula (14) to update the fire line coordinate set in the wildfire spread state vector.

[0083] (14) In formula (14), Is the fire line coordinate set in the updated wildfire spread state vector, that is, the predicted Fire line coordinate set at a moment, Is the length of the time step, that is, the time difference between two adjacent time steps, for example, it is The time difference between a moment and A moment.

[0084] Is a two-dimensional unit vector. The velocity on the two-dimensional plane can be decomposed into The velocity in the direction of The speed in the direction can be updated The coordinates in the direction can be updated direction coordinates. Based on this principle, the above formula (14) can be expanded into the forms of formula (15) and formula (16).

[0085] For the set of fireline coordinates at any one coordinate , the coordinate is updated (predicted) at time to be , where is represented by formula (15), is represented by formula (16).

[0086] (15) (16) In formula (15) and formula (16), represents the coordinate after coordinate update, that is, the coordinate position at time. The meanings of other parameters in formula (15) and formula (16) are the same as those in formula (11) and formula (12), and are omitted here for detailed description.

[0087] When the stop condition is met, step 204 can be stopped. The stop conditions include but are not limited to: Reaching the preset maximum time step, the fire spread speed is less than the preset speed threshold and the real-time fire intensity of all points on the fireline is less than the preset fire intensity threshold (indicating that the combustion is approaching to stop and the update can be stopped), and all fireline coordinates have reached the preset geographical boundary.

[0088] In step 205, a wildfire report is generated based on the updated set of fireline coordinates.

[0089] The wildfire report also includes the real-time fire intensity of the wildfire, and the real-time fire intensity is calculated based on the set of environmental parameters.

[0090] The definition of the real-time fire intensity (Fireline Intensity, FI) is: the heat released per unit time and per unit length of the fireline. Optionally, the real-time fire intensity is calculated using formula (17).

[0091] (17) In formula (17), is The real-time fire intensity at is the calorific value of combustibles per unit mass at is the load of combustibles per unit area at 2 ). is the fire front spread speed at , which is calculated by formula (11).

[0092] The calorific value of combustibles per unit mass has a corresponding relationship with the vegetation type , and this corresponding relationship can be pre-stored in a computer device. Substituting the real-time into this corresponding relationship can obtain the real-time at .

[0093] Optionally, the load of combustibles per unit area at is calculated by formula (18).

[0094] (18) In formula (18), is the basic combustible load at , is the weight. The meanings of other parameters in formula (18) are the same as those in formula (7) and formula (17), and are not elaborated here.

[0095] Among them, the basic combustible load at is calculated by formula (19).

[0096] (19) In formula (19), , are the weights. The meanings of other parameters in formula (19) are the same as those in formula (6) and formula (18), and are not elaborated here.

[0097] Through formulas (17) to (19), it is possible to obtain the real-time fire intensity in the first area.

[0098] The wildfire report can include the updated set of fire line coordinates, the real-time fire intensity distribution of each fire line coordinate, etc. It is also possible to further process the updated set of fire line coordinates to generate a spread path map, calculate the area of the region affected by the wildfire, etc.

[0099] In step 206, correct the updated fire line coordinates set and real-time fire intensity in the wildfire report.

[0100] Ideally, the predicted fire line coordinates can be achieved according to formulas (14) to (16), and the real-time fire intensity can be calculated through formulas (17) to (19). However, during the actual combustion process of wildfires, there are still many factors that will affect the predicted fire line coordinates and real-time fire intensity. Therefore, it is necessary to correct the predicted fire line coordinates and real-time fire intensity.

[0101] There is a complex interaction relationship between wildfires and the environment. Wildfires are not only affected by environmental factors, but also significantly change the surrounding environmental conditions, thereby affecting their own spreading behavior. Therefore, this dynamic two-way coupling relationship needs to be considered in wildfire spread simulation, especially in large-scale and long-duration wildfire events, the influence of this factor will be more critical. That is to say, wildfires will probably have an impact on the environment.

[0102] The impact of wildfires on the environment will cause errors in the parameters calculated in the environmental feature set. This kind of error is mainly considered from the following three aspects: the error of wind speed and wind direction angle, the error of temperature, and the error of fuel load. The error of wind speed and wind direction angle and the error of temperature will affect the accuracy of the predicted fire line coordinates; the error of fuel load will affect the accuracy of real-time fire intensity.

[0103] A large amount of heat energy released during the wildfire combustion process will change the local atmospheric environment, generating the "fire meteorology" phenomenon. The high temperature generated by wildfires will cause the surrounding air to heat up rapidly, forming strong thermal convection, which will change the local wind field structure and form a converging air current towards the fire source center. This self-generated wind field is often stronger than the original environmental wind, which can further intensify the fire, especially in large wildfires, the "fire whirlwind" phenomenon may also occur. Therefore, wildfire combustion will cause errors in wind speed and wind direction angle.

[0104] During the wildfire combustion process, a large amount of oxygen will be consumed and substances such as carbon dioxide, water vapor and soot will be released, changing the local atmospheric composition. Especially the generation of soot will significantly reduce the solar radiation intensity and affect the surface temperature distribution. At the same time, the smoke generated by the fire will reduce visibility and affect fire field observation and rescue operations. Therefore, wildfire combustion will cause temperature errors.

[0105] Wildfires will cause a sharp change in the surface vegetation cover. After burning, the fuel in the area decreases and the surface albedo changes, which will further affect the local radiation balance and energy exchange process. Therefore, wildfire combustion will cause errors in fuel load.

[0106] The above three aspects can be corrected respectively using the following formulas (20) to (24).

[0107] Optionally, the wind speed and the set of wind direction angles at are corrected using Equation (20). .

[0108] (20) In Equation (20), is the corrected ; is the fire-induced wind field coefficient, which can be obtained through wind tunnel experiments, CFD (Computational Fluid Dynamics) simulations, or regression analysis of historical fire data, and is usually related to the fire intensity and local terrain. is the distance from to the fire line, and is a characteristic distance parameter, which is a learnable value. is the unit vector pointing to the fire origin

[0109] (21) In Equation (21), is the distance between and the fire origin

[0110] Optionally, the temperature at is corrected using Equation (22). .

[0111] (22) In Equation (22), is the corrected , is the temperature at is the temperature decay coefficient, which can be obtained by regression fitting based on measured data (such as the surface temperature change in the fire area), or by calculating the shielding effect of smoke on solar radiation using an atmospheric transmission model. . is the smoke concentration at

[0112] Optionally, the smoke concentration at It can be estimated based on remote sensing data, and this process can be expressed by formula (23).

[0113] (23) In formula (23), is the aerosol optical depth data collected by the remote sensing satellite (such as MODIS, VIIRS, Sentinel-5P, etc.) at is the empirical regression coefficient, The unit of 3 is kg / m

[0114] Optionally, the fuel load at is corrected using formula (24). .

[0115] (24) In formula (24), is the corrected , is the fuel consumption coefficient, which can be determined through combustion experiments (such as the combustion rates of different vegetation types) , and can also be calculated using existing combustion models (such as the fuel loss in the Rothermel fire behavior model) . The meanings of other parameters in formula (24) are the same as those in formula (17) and formula (22), and are not elaborated here.

[0116] When correcting the parameters in the above three aspects, the model can be further refined according to the actual situation. For example, in a large-scale wildfire that lasts for a long time, the heat generated by the fire may even affect the local weather system, induce the formation of thunderstorm clouds, and then may lead to rainfall or trigger new ignition points. This "fire-meteorology" coupling effect may form a self-reinforcing positive feedback loop, making it difficult to control the fire. For such a more complex situation, a more advanced atmosphere-fire coupling model needs to be introduced for simulation during specific implementation, considering the impact of fire heat release on the atmospheric boundary layer structure and the possible precipitation feedback effect.

[0117] Generally, after correcting the parameters at time , the corrected parameters should be substituted into the operation at time

[0118] In the case where the area of the study area is large and the first area cannot be regarded as a plane, longitude and latitude coordinates need to be used, and then the longitude and latitude coordinates are transformed to obtain projected coordinates Perform calculations later. When implementing, for example, the UTM (Universal Transverse Mercator Grid System) projection can be used to perform coordinate conversion on longitude and latitude coordinates.

[0119] In the case of complex climate and terrain in the first region, the complex climate and terrain will affect the wildfire burning in the first region. At this time, it is also necessary to correct the impact of complex climate and terrain on the wildfire. Among them, complex climate and terrain include, but are not limited to: large diurnal temperature difference, rainfall, canyons, etc.

[0120] Parameters such as temperature and relative humidity are not updated in real time, but are updated with meteorological forecast data. Under normal circumstances, it may be updated hourly. Then, in the case of a large diurnal temperature difference, there may be large fluctuations in temperature and relative humidity within two adjacent hours. That is to say, the temperature and relative humidity at each time step are unclear, which will affect the predicted set of fire line coordinates. Therefore, in this case, it is necessary to correct the temperature and relative humidity.

[0121] Rainfall has an obvious inhibitory effect on wildfires. Rainfall will cause a decrease in the real-time fire intensity and also affect the spread speed of the fire line, thereby affecting the predicted set of fire line coordinates and the real-time fire intensity.

[0122] The wildfire spread situation under complex terrain conditions is more complex. Taking the wildfire spread in a canyon as an example, due to the canyon effect, the wind speed in the canyon is often large, resulting in errors in wind speed and wind direction angle, and further affecting the predicted set of fire line coordinates.

[0123] The following formulas (25) to (28) can be used to correct the situation of complex climate and terrain.

[0124] Optionally, in the case of a large diurnal temperature difference, use the daily temperature change model to determine the temperature at time. This process can be represented by formula (25).

[0125] (25) In formula (25), is the daily temperature change model, indicating the temperature at time; is the daily average temperature, is the amplitude of the temperature in a day, is the time point when the highest temperature in a day appears, is the current time. In addition, when calculating formula (25), 1 day needs to be converted to the same unit as and For example, in and When the unit is seconds, 1 day needs to be converted to 86,400 seconds before calculation.

[0126] Optionally, in the case of large day-night temperature differences, use formula (26) to calculate the relative humidity at a certain moment.

[0127] (26) In formula (26), represents the relative humidity at a certain moment, is the reference humidity of the relative humidity, is the temperature-humidity relationship coefficient. The meanings of other parameters in formula (26) are the same as those in formula (25) and are not elaborated here.

[0128] Optionally, in the case where there is a rainfall area in the first region, use formula (27) to correct the fire intensity within the rainfall area.

[0129] (27) In formula (27), represents the corrected real-time fire intensity when it is located within the rainfall area, represents the real-time fire intensity when it is located within the rainfall area, which is calculated using formula (17). is the rainfall influence coefficient, represents the rainfall intensity at a certain location.

[0130] Optionally, in the case where there is a canyon terrain in the first region, use formula (28) to correct the wind speed and wind direction angle within the canyon terrain.

[0131] (28) In formula (28), is the corrected (here is located within the canyon terrain of the first region), is the canyon effect coefficient, is the canyon depth, is the canyon width. The meanings of other parameters in formula (28) are the same as those in formula (1) and are not elaborated here.

[0132] For other special cases, corrections can also be made according to the actual situation.

[0133] In actual situations, humans can also have an impact on wildfires. For example, firefighting operations can suppress the intensity of the fire. In this case, the corrected real-time fire intensity can be expressed by formula (29).

[0134] (29) In formula (29), is the corrected (here is within the area affected by firefighting operations in the first region), is the fire extinguishing efficiency coefficient, representing the fire extinguishing efficiency, The larger it is, the higher the fire extinguishing efficiency. represents the firefighting force invested, represents the distance from the location to the fire line.

[0135] The weights and coefficients involved in the above formulas (1) to (29) (such as to , , etc.) can take empirical values or be obtained using deep learning methods. If the weights are obtained based on empirical values, the sources of empirical data can include historical fire case data, field test data, expert empirical values, etc. Statistical regression analysis, parameter sensitivity analysis, Monte Carlo simulation, etc. can be used to analyze the empirical data to obtain empirical values.

[0136] If deep learning methods are used to obtain the weights, a model can be selected for establishment. This model includes CNN, LSTM, and Transformer. Among them, CNN is used to process spatial features, LSTM is used to process temporal features, and Transformer is used to process multimodal data. The model is trained by inputting multispectral remote sensing images, DEM data, meteorological data, historical fire data, etc. into the model, and finally the weights are obtained based on the trained model.

[0137] In the actual application process, empirical values and deep learning methods can be combined for use. For example, empirical values can be used initially to quickly establish a model, and deep learning methods can be gradually introduced as data accumulates. The two methods are combined and verified with each other, and the parameters are updated and optimized regularly. It is also necessary to pay attention to establishing a parameter verification mechanism, considering the adaptability of different regions, and attaching importance to the real-time performance of the model in wildfire disaster emergency applications.

[0138] The following are the device embodiments of this application. For the details not described in detail in the device embodiments, reference can be made to the above method embodiments.

[0139] Figure 3The structural schematic diagram of the wildfire spread prediction device provided by an exemplary embodiment of the present disclosure is shown. Refer to Figure 3 , the wildfire spread prediction device 300 includes: a first acquisition module 301, a second acquisition module 302, a state vector construction module 303, and an update module 304.

[0140] The first acquisition module 301 is configured to acquire an environmental feature set of the first area based on the remote sensing image of the first area, where the first area is an area with a wildfire, and the environmental feature set includes elevation data, slope, aspect, vegetation type, water content of combustibles, wind speed, wind direction angle, temperature, and relative humidity; The second acquisition module 302 is configured to acquire a fire source feature vector of the wildfire in the first area based on the environmental feature set, and the fire source feature vector includes the first fire intensity at the ignition moment; The state vector construction module 303 is configured to construct a wildfire spread state vector based on the environmental feature vector and the fire source feature vector. The wildfire spread state vector includes a set of fire line coordinates, a fire line spread speed, and a fire line spread direction, and the fire line spread speed is determined based on the first fire intensity; The update module 304 is configured to update the set of fire line coordinates in the wildfire spread state vector based on the fire line spread speed and the fire line spread direction, so as to predict the wildfire in the first area.

[0141] Optionally, the update module 304 is further configured to update the set of fire line coordinates in the wildfire spread state vector by using the following formula:

[0142] where, is the set of fire line coordinates in the updated wildfire spread state vector, representing the set of fire line coordinates at time is the length of the time step, representing the time difference between time and time

[0143] Optionally, the device further includes: a generation module 305 and a correction module 306.

[0144] The generation module 305 is configured to generate a wildfire report based on the updated set of fire line coordinates, and the wildfire report further includes the real-time fire intensity of the wildfire, and the real-time fire intensity is calculated based on the environmental parameter set.

[0145] The correction module 306 is configured to correct the updated set of fire line coordinates and the real-time fire intensity in the wildfire report.

[0146] It should be noted that when the wildfire spread prediction device provided in the above embodiments predicts the wildfire spread, only the division of the above functional modules is used for illustration. In actual applications, the above functions can be assigned to different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the wildfire spread prediction device provided in the above embodiments and the embodiments of the wildfire spread prediction method belong to the same concept. For the specific implementation process, please refer to the method embodiments and will not be elaborated here.

[0147] The division of modules in the embodiments of the present disclosure is illustrative. It is only a logical function division. In actual implementation, there may be other division methods. In addition, in each embodiment of the present disclosure, the functional modules can be integrated in a processor, or exist independently physically, or two or more modules can be integrated into one module. The above integrated modules can be implemented in the form of hardware or in the form of software functional modules.

[0148] If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present disclosure, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions to enable a terminal device (which can be a personal computer, a mobile phone, or a communication device, etc.) or a processor to execute all or part of the steps of the method in each embodiment of the present disclosure. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks, or optical discs that can store program codes.

[0149] Figure 4 is a schematic structural diagram of the computer device provided in the embodiments of the present disclosure. As Figure 4 shown, the computer device 400 includes: a processor 401 and a memory 402.

[0150] The processor 401 may include one or more processing cores, such as a quad-core processor, an octa-core processor, etc. The processor 401 may be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), or PLA (Programmable Logic Array). The processor 401 may also include a main processor and a coprocessor. The main processor is a processor used to process data in the wake state, also known as the CPU (Central Processing Unit); the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, the processor 401 may be integrated with a GPU (Graphics Processing Unit), and the GPU is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 401 may further include an AI (Artificial Intelligence) processor, and the AI processor is used to process computational operations related to machine learning.

[0151] The memory 402 may include one or more computer-readable storage media, and the computer-readable storage media may be non-transitory. The memory 402 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices and flash storage devices. In some embodiments, the non-transitory computer-readable storage media in the memory 402 is used to store at least one instruction, and the at least one instruction is used to be executed by the processor 401 to implement the wildfire spread prediction method provided in the embodiments of the present disclosure.

[0152] Those skilled in the art can understand that Figure 4 the structure shown in

[0153] does not constitute a limitation on the computer device 400, and may include more or fewer components than shown in the figure, or combine certain components, or adopt a different component arrangement.

[0154] The embodiments of the present disclosure also provide a non-temporary computer-readable storage medium. When the instructions in the storage medium are executed by the processor of the computer device, the computer device can execute the wildfire spread prediction method provided in the embodiments of the present disclosure.

[0155] The above are only alternative embodiments of the present disclosure and are not intended to limit the present disclosure. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present disclosure shall be included within the protection scope of the present disclosure.

Claims

1. A method for predicting the spread of wildfires, characterized in that: The method comprises: Based on a remote sensing image of a first area, an environmental feature set of the first area is obtained, where the first area is an area where a wildfire exists, and the environmental feature set includes elevation data, slope, slope aspect, vegetation type, moisture content of combustibles, wind speed, wind direction angle, temperature, and relative humidity; Based on the environmental feature set, obtaining a fire source feature vector of the wildfire in the first area, wherein the fire source feature vector includes a first fire intensity at the time of fire ignition; Based on the environmental feature vector and the fire source feature vector, construct a wildfire spread state vector, wherein the wildfire spread state vector includes a fire line coordinate set, a fire line spread speed, and a fire line spread direction, wherein the fire line spread speed is determined based on the first fire intensity; Based on the fire line spread speed and the fire line spread direction, the fire line coordinate set in the wildfire spread state vector is updated to predict the wildfire in the first area.

2. The method according to claim 1, characterized in that The wildfire spread state vector is expressed by the following formula: in, for The wildfire spread state vector at time , for The fire line coordinate set at time, for The line of fire coordinates set at the moment The fire line propagation speed of each fire line coordinate in, for The line of fire coordinates set at the moment The fire spread direction of each fire coordinate in .

3. The method according to claim 2, characterized in that express The line of fire coordinates set at the moment Any line of fire coordinates in , Fire line coordinates at the moment Fire spread speed It is calculated using the following formula: Fire line coordinates at the moment The direction of fire spread It is calculated using the following formula: in, for Always in the line of fire The speed of fire spread at for Always in the line of fire The slope at for Always in the line of fire The wind speed at for Always in the line of fire The wind direction angle at for Always in the line of fire The slope direction at for Always in the line of fire The normalized water index at for Always in the line of fire The relative humidity at for Always in the line of fire The moisture content of the combustible at , , , , , , is the weight.

4. The method according to claim 2 or 3, characterized in that The updating of the fire line coordinate set in the wildfire spread state vector comprises: The following formula is used to update the fire line coordinate set in the wildfire spread state vector: in, is the fire line coordinate set in the updated wildfire spread state vector, indicating The fire line coordinate set at time, is the length of the time step, indicating Moment and The time difference between moments.

5. The method according to any one of claims 1 to 3, characterized in that: The method further comprises: A wildfire report is generated based on the updated fire line coordinate set, wherein the wildfire report also includes the real-time fire intensity of the wildfire, and the real-time fire intensity is calculated based on the environmental parameter set.

6. The method according to claim 5, characterized in that The method further comprises: The updated fire line coordinate set and the real-time fire intensity in the wildfire report are corrected.

7. A wildfire spread prediction device, characterized in that: The device comprises: A first acquisition module is used to acquire an environmental feature set of a first area based on a remote sensing image of the first area, wherein the first area is an area where a wildfire exists, and the environmental feature set includes elevation data, slope, slope aspect, vegetation type, moisture content of combustibles, wind speed, wind direction angle, temperature, and relative humidity; A second acquisition module, configured to acquire a fire source feature vector of a wildfire in the first area based on the environmental feature set, wherein the fire source feature vector includes a first fire intensity at a fire ignition moment; a state vector construction module, configured to construct a wildfire spread state vector based on the environment feature vector and the fire source feature vector, wherein the wildfire spread state vector includes a fire line coordinate set, a fire line spread speed, and a fire line spread direction, wherein the fire line spread speed is determined based on the first fire intensity; An updating module is used to update the fire line coordinate set in the wildfire spread state vector based on the fire line spread speed and the fire line spread direction, so as to predict the wildfire in the first area.

8. A computer device, characterized in that: The computer device comprises: a memory and a processor, wherein at least one computer program is stored in the memory, and the at least one computer program is loaded and executed by the processor to implement the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores at least one computer program, and the at least one computer program is loaded and executed by a processor to implement the method according to any one of claims 1 to 6.

10. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images