Control method and system suitable for unmanned aerial vehicle to spray flame retardant

Abstract

The invention discloses a control method and system suitable for an unmanned aerial vehicle to spray a flame retardant, and relates to the technical field of unmanned aerial vehicles, the method comprises the following steps: carrying out spectral analysis on a fire area image, and determining a fire source point according to a spectral analysis result; determining an initial flame retardant spraying area according to the fire area, correcting the initial flame retardant spraying area by considering the position of a fire source point, a wind power factor, a terrain factor and a vegetation factor to obtain a flame retardant spraying area, and generating a flame retardant spraying strategy; and controlling the unmanned aerial vehicle to execute a corresponding flame retardant spraying strategy in the flame retardant spraying area. The fire source point is determined through spectral analysis, various dynamic factors such as wind power, terrain and vegetation are considered, the spraying area of the flame retardant can be intelligently adjusted, the corresponding spraying strategy is automatically generated and executed, the flame retardant effect is improved, and meanwhile unnecessary resource waste is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle control, and in particular to a control method and system suitable for spraying flame retardants on unmanned aerial vehicles. Background Art

[0002] In recent years, forest fires have become more and more severe. In the past, fire control was usually carried out by manually spraying flame retardants on the fire area. However, the manual method would threaten the safety of firefighters.

[0003] The patent (CN 112915421 B) discloses a method for spraying flame retardants for forest fires based on drones, which includes the following steps: using an infrared temperature measurement component to detect the fire area, allowing the GPS positioning component to obtain the first position information of the fire area, and sending the first position information to the remote control console; obtaining the first wind direction information of the fire area through a wind direction sensor, and sending the first wind direction information to the remote control console; determining the spraying area of ​​the flame retardant, the spraying area is the area around the fire area that has not caught fire; controlling the flame retardant spraying component to start spraying the flame retardant to the spraying area through the remote control console to form a flame retardant isolation zone. The invention uses a method of spraying flame retardants by drones to control the fire, which can ensure the safety of personnel. In addition, the flame retardant isolation zone formed by spraying flame retardants around the fire area spreads.

[0004] Although the above method realizes the automatic spraying of flame retardants, it is unable to determine the type of flame retardants to be sprayed and the spraying area according to the actual fire situation. Summary of the invention

[0005] In view of the defects in the prior art, the present invention provides a control method and system suitable for spraying flame retardants by unmanned aerial vehicles.

[0006] On the one hand, an embodiment of the present invention provides a control method for spraying flame retardant by a drone, the method comprising: Acquire an infrared image and an optical image of the fire scene, fuse the infrared image and the optical image to obtain a fused image, and identify a fire area image from the fused image; Performing spectrum analysis on the image of the fire area, and determining the fire source point according to the spectrum analysis result; Determine an initial flame retardant spraying area according to the fire area, modify the initial flame retardant spraying area by taking into account the location of the fire source, wind factor, terrain factor and vegetation factor, obtain the flame retardant spraying area, and generate a flame retardant spraying strategy; The UAV is controlled to execute a corresponding flame retardant spraying strategy in the flame retardant spraying area.

[0007] Preferably, acquiring an infrared image and an optical image of the fire scene, and fusing the infrared image and the optical image to obtain a fused image comprises: Preprocessing the infrared image and the optical image; Using a first coarse feature extraction module and a first attention module to extract features from the preprocessed optical image to obtain a first feature map; Using a second coarse feature extraction module and a second attention module to extract features from the preprocessed optical image to obtain a second feature map; The first feature map and the second feature map are fused to obtain a fused image.

[0008] Preferably, performing spectrum analysis on the fire area image, and determining the fire source point according to the spectrum analysis result comprises: Extracting a flame region image from the fused image, performing Fourier transform on the flame region image to obtain a frequency domain representation; Calculate the total energy of each pixel in the fused image in the key spectrum band according to the frequency domain representation, calculate the average energy of a reference point far away from the fire area in the key spectrum band, and use the average energy as a benchmark to calculate the energy ratio between the pixel and the reference point; When the energy ratio is greater than a set value, the pixel point is determined to be a fire source point, and the size of the fire is determined based on the difference between all energy ratios and the set value.

[0009] Preferably, the initial flame retardant spraying area is corrected by taking into account the location of the fire source, wind factor, terrain factor and vegetation factor to obtain the flame retardant spraying area, including: Discretize the regional edge of the fire area into a number of edge fire points, and determine the fire source spread amount according to the position information of the edge fire points and the position information of the fire source point, wherein the fire source spread amount includes a spread direction and a spread distance; Calculate the difference between the energy ratio of all pixels on the edge fire point and the fire source point and a set value, and determine the initial fire intensity according to the fire source spread and the difference; The initial fire intensity is corrected by taking into account wind factor, terrain factor and vegetation factor to obtain a target fire intensity; Determine edge points of the flame retardant spraying area according to the target fire intensity, connect the edge points of the flame retardant spraying area to obtain a flame retardant spraying edge, and combine the flame retardant spraying edge and the initial flame retardant spraying area to obtain a flame retardant spraying area.

[0010] Preferably, the regional edge of the fire area is discretized into a number of edge fire points, including: determining the discreteness of the edge fire points according to the smoothness of the line change of the regional edge; and discretizing the regional edge into a number of edge fire points according to the discreteness.

[0011] Preferably, the initial fire intensity is corrected by taking into account wind factor, terrain factor and vegetation factor, and the formula for obtaining the target fire intensity is: in, Indicates the target fire. Indicates the initial fire. represents the comprehensive correction coefficient, represents the wind factor function, represents the terrain factor function, represents the vegetation factor function, Represents the wind speed vector at the fire source, including the magnitude and direction of the wind speed. It indicates the angle between wind direction and the direction of fire spread. The slope indicates the direction of fire spread. It represents the angle between the slope and the direction of fire spread. Indicates the vegetation type, Indicates vegetation density.

[0012] Preferably, the generated flame retardant spraying strategy includes: determining the type of flame retardant to be sprayed according to the target fire intensity, planning the spraying path of the drone using an optimization algorithm, and generating the flame retardant spraying strategy.

[0013] On the other hand, a control system suitable for spraying flame retardant by a drone, the system comprising: An image recognition module is used to obtain an infrared image and an optical image of the fire scene, fuse the infrared image and the optical image to obtain a fused image, and recognize the fire area image from the fused image; A preliminary analysis module is used to perform spectrum analysis on the image of the fire area and determine the fire source and fire intensity according to the spectrum analysis results; A strategy determination module is used to determine the initial flame retardant spraying area according to the fire area and the fire source point, divide the fire area and the flame retardant spraying area into a number of fire sub-areas and a number of flame retardant spraying areas with the fire source point as the center, and correct the fire intensity of the fire sub-area by considering the terrain factor, vegetation factor and wind factor of the fire sub-area to obtain the regional fire intensity, and correct the initial flame retardant spraying area according to the regional fire intensity to obtain the flame retardant spraying area, and generate a flame retardant spraying strategy; The strategy execution module controls the UAV to execute the corresponding flame retardant spraying strategy in the flame retardant spraying area.

[0014] The beneficial effects of the present invention are as follows: the present invention not only determines the initial flame retardant spraying area based on the fire area, but also comprehensively considers the influence of dynamic factors such as wind, terrain and vegetation on the fire. Through sophisticated algorithm correction, a more accurate flame retardant spraying area is obtained, and the fire level of the fire area can be evaluated according to the obtained target fire level. A spraying strategy including flame retardant type and spraying amount is automatically generated according to the fire level, which improves the flame retardant effect and reduces unnecessary waste of resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for the specific embodiments or the description of the prior art. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn according to the actual scale.

[0016] Figure 1 A flow chart of a control method for spraying flame retardant on a drone is provided for an embodiment of the present invention; Figure 2 A system block diagram of a control system suitable for spraying flame retardants by a drone is provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The following embodiments of the technical solution of the present invention are described in detail in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and are therefore only used as examples, and cannot be used to limit the protection scope of the present invention. It should be noted that, unless otherwise specified, the technical terms or scientific terms used in this application should have the common meanings understood by those skilled in the art to which the present invention belongs.

[0018] Example 1 like Figure 1 As shown, an embodiment of the present invention provides a control method for spraying flame retardants on a drone, the method comprising: Step 1: Acquire an infrared image and an optical image of the fire scene, fuse the infrared image and the optical image to obtain a fused image, and identify the fire area image from the fused image.

[0019] In this embodiment of the present invention, step 1 includes: Step 11, preprocessing the infrared image and the optical image.

[0020] Specifically, select a drone model with high stability and load capacity to ensure that it can carry the required sensors and spraying devices. The drone is equipped with a high-resolution infrared camera and an optical camera to ensure that detailed image information of the fire scene can be captured at the same time. The infrared camera is used to capture the thermal radiation information of the fire source, and the optical camera provides clear visible light images. Set the drone's flight altitude and speed to ensure that the clarity and coverage of the image capture meet the needs of subsequent processing, and start the infrared and optical cameras simultaneously to continuously shoot the fire scene and obtain infrared image data and optical image data.

[0021] More specifically, preprocessing includes denoising the original optical and infrared images, using methods such as Gaussian filtering or median filtering to remove noise interference in the image. Histogram equalization is performed to enhance the contrast and brightness of the image to make the flame area more prominent. The image is cropped and scaled to ensure that the processed area matches the actual size of the fire scene.

[0022] Step 12, using a first coarse feature extraction module and a first attention module to extract features from the preprocessed optical image to obtain a first feature map; Specifically, a convolutional neural network (CNN) is used as a coarse feature extraction module to extract features from the preprocessed infrared and optical images. CNN can extract features such as edges, textures, and shapes in images through multi-layer convolution and pooling operations.

[0023] Step 13: Use the second coarse feature extraction module and the second attention module to extract features from the preprocessed optical image to obtain a second feature map.

[0024] Specifically, attention mechanisms such as SE (Squeeze-and-Excitation) module or CBAM (Convolutional Block Attention Module) are introduced to enhance the focus on key features (such as flame edges and hot spots). Through the attention mechanism, the network can adaptively adjust the weight of the feature map to improve the accuracy and robustness of feature extraction.

[0025] Step 14: Fusing the first feature map and the second feature map to obtain a fused image.

[0026] Specifically, a weighted average fusion strategy is used to fuse the infrared feature map and the optical feature map. According to the performance characteristics of the flame in the infrared and optical images, a reasonable weight coefficient is set so that the fused image contains both the thermal radiation information of the flame and its visible light characteristics.

[0027] More specifically, the fire area image is identified from the fused image using existing technology, which will not be described in detail in this embodiment. The fire area image is used to indicate the fire area.

[0028] Step 2: Perform spectrum analysis on the fire area image and determine the fire source point according to the spectrum analysis result.

[0029] In an embodiment of the present invention, step 2 includes: step 21, extracting a flame area image from the fused image, performing Fourier transform on the flame area image, and obtaining a frequency domain representation; step 22, calculating the total energy of each pixel point in the fused image in the key spectrum band according to the frequency domain representation, calculating the average energy of a reference point far away from the fire area in the key spectrum band, and using the average energy as a benchmark to calculate the energy ratio of the pixel point to the reference point; step 23, when the energy ratio is greater than a set value, determining the pixel point as a fire source point, and determining the size of the fire according to the difference between all energy ratios and the set value.

[0030] Specifically, the flame area in the fused image is subjected to Fourier transform to convert the image from the spatial domain to the frequency domain. Fourier transform can reveal the frequency components in the image, and the flame area is usually manifested as high-frequency components in the frequency domain. According to the characteristics of the flame in the frequency domain, the key spectrum band is selected for energy calculation. The key spectrum band can be determined based on experiments or experience, and usually contains the main frequency components of the flame. The total energy of each pixel in the fused image in the key spectrum band is calculated to obtain an energy map. The energy map reflects the intensity distribution of the flame area. A reference point far away from the fire area is selected to calculate its average energy in the key spectrum band. The average energy is used as a benchmark to calculate the energy ratio of each pixel to the reference point. An energy ratio threshold is set to screen out pixels with an energy ratio greater than the threshold as fire source points.

[0031] At the same time, the difference between the energy ratio and the threshold can be used to divide the fire intensity and provide a basis for subsequent spraying strategies.

[0032] Step 3, determine the initial flame retardant spraying area according to the fire area, modify the initial flame retardant spraying area by considering the location of the fire source, wind factor, terrain factor and vegetation factor, obtain the flame retardant spraying area, and generate a flame retardant spraying strategy.

[0033] In this embodiment of the present invention, step 3 includes: Step 31, discretize the regional edge of the fire area into a number of edge fire points, and determine the fire source spread amount according to the position information of the edge fire points and the position information of the fire source point, wherein the fire source spread amount includes the spread direction and the spread distance.

[0034] It should be noted that discretizing the regional edge of the fire area into a number of edge fire points includes: determining the discreteness of the edge fire points according to the smoothness of the line change of the regional edge; and discretizing the regional edge into a number of edge fire points according to the discreteness.

[0035] Specifically, first, edge lines are identified from the image of the fire area obtained in step 1, and then the curvature or slope change of the line is calculated to quantify the smoothness of the line. For example, a sliding window can be set to slide on the edge line, and the slope change rate of the line in each window can be calculated. The area with a larger slope change rate indicates that the line changes violently; while the area with a smaller slope change rate indicates that the line changes gently, the greater the discreteness of the edge fire point, the discreteness of the edge fire point can be determined according to the gentleness of the edge line. The discreteness can be understood as the distance between adjacent edge fire points. In the area where the line changes violently, since there may be multiple branch points for the spread of fire, the discreteness of the edge fire point should be set smaller to more accurately locate these perceived fire changes. On the contrary, in the area where the line changes gently, since the spread of fire may be relatively uniform, the discreteness of the edge fire point can be set larger. According to the determined discreteness, the regional edge of the fire area is discretized into several edge fire points. This can be achieved by sampling on the edge line according to the set discreteness. Each sampling point can be considered as a potential edge fire point, whose location is determined by the coordinates of the sampling point.

[0036] Step 32, calculating the difference between the energy ratio corresponding to all pixel points on the edge fire point and the fire source point and a set value, and determining the initial fire intensity according to the fire source spread and the difference.

[0037] Specifically, the product of the fire source spread and the sum of all differences represents the initial fire intensity, the specific direction and size attributes of the initial fire intensity.

[0038] Step 33, taking into account wind factor, terrain factor and vegetation factor, the initial fire intensity is corrected to obtain a target fire intensity.

[0039] Wind usually increases the speed of fire spread and changes the direction of spread. The wind factor can be determined based on wind speed and wind direction, and is usually a vector proportional to wind speed and related to wind direction. Topography such as slope and aspect can affect the spread of fire. Going uphill usually speeds up the fire, while going downhill may slow it down. The type, density and moisture of vegetation can affect the spread of fire. Dry and flammable vegetation will accelerate the spread of fire. Therefore, when determining the fire, it is necessary to consider not only the static image, but also the impact of dynamic factors on the fire.

[0040] Specifically, the initial fire intensity is corrected by taking into account wind factor, terrain factor and vegetation factor, and the formula for obtaining the target fire intensity is: in, Indicates the target fire. Indicates the initial fire. represents the comprehensive correction coefficient, represents the wind factor function, represents the terrain factor function, represents the vegetation factor function, Represents the wind speed vector at the fire source, including the magnitude and direction of the wind speed. It indicates the angle between wind direction and the direction of fire spread. The slope indicates the direction of fire spread. It represents the angle between the slope and the direction of fire spread. Indicates the vegetation type, Indicates vegetation density.

[0041] Among them, wind speed and wind direction are obtained by measuring the data at the fire source in real time through meteorological sensors (such as wind speed and wind direction sensors). Ultrasonic wind speed and wind direction sensors use the Doppler effect to measure wind speed and wind direction, and can provide high-precision and high-stability data. Wind speed vector Based on the measured wind speed and wind direction, the wind speed vector can be determined, which contains the magnitude and direction information of the wind speed. The angle between the wind direction and the direction of fire spread can be calculated to evaluate the impact of wind on the direction of fire spread. Slope uses topographic maps or DEM (digital elevation model) data to extract slope information through GIS (geographic information system) or related software (such as ArcGIS) surface analysis tools. Slope direction is also based on topographic maps or DEM data, and the slope direction information is extracted through GIS software. The angle between the slope direction and the direction of fire spread is calculated to evaluate the impact of terrain on the speed of fire spread. Vegetation type is identified by remote sensing technology (such as RS) or determined in combination with field survey data. Vegetation density is obtained through interpretation of remote sensing images or on-site survey measurements. UAVs can retrieve weather data, vegetation databases, and terrain databases to obtain or predict data. It should be noted that the calculation of factors such as terrain and vegetation type in the calculation formula of the target fire intensity is the average of all slopes or vegetation types on the straight line from the fire source to the edge fire point. The function in the formula is determined based on experiments or experience, and can also be fitted by machine learning algorithms.

[0042] Step 34, determining edge points of the flame retardant spraying area according to the target fire intensity, connecting the edge points of the flame retardant spraying area to obtain a flame retardant spraying edge, and combining the flame retardant spraying edge and the initial flame retardant spraying area to obtain a flame retardant spraying area.

[0043] Specifically, the target fire intensity can characterize the approximate fire intensity of each edge fire point. The distance between the edge fire point and the edge point of the flame retardant spraying area is determined according to the size of the target fire. Taking the edge fire point as the starting point, the distance is extended along the direction of the target fire to obtain the edge point of the flame retardant spraying area.

[0044] Step 35, calculate the mean of the target fire intensity of all edge fire points, determine the fire intensity of the fire area according to the mean, determine the type of flame retardant according to the fire intensity, use the optimization algorithm to plan the spraying path of the drone, and generate a flame retardant spraying strategy.

[0045] It should be noted that the embodiment of the present invention constructs a fire level-flame retardant type relationship library. Of course, a fire level-vegetation flammability level-ambient temperature-ambient suitability-flame retardant type relationship library can be constructed. A fire level-vegetation type-ambient condition-flame retardant type relationship library can also be constructed. It can be constructed according to actual needs and is not limited in this embodiment.

[0046] Specifically, the mean of all edge-fired target fires is calculated, and the fire intensity of the fire area is determined according to the mean. The type of flame retardant is determined according to the fire intensity, vegetation type and environmental conditions. For example, for plant fiber materials, flame retardants that can effectively penetrate into plant fibers and are well compatible with them should be selected. For example, some organic phosphorus flame retardants or nitrogen flame retardants; for wood, organic boron flame retardants (such as zinc borate) should be selected because of their good heat resistance, low toxicity and smoke suppression, which are suitable for wood flame retardant. In addition, nitrogen flame retardants such as melamine and dicyandiamide can also be used for wood flame retardant. For small fire sources, some fast-responding flame retardants can be selected, such as some intumescent flame retardants. This type of flame retardant can quickly foam and expand when heated to form a protective layer, thereby quickly preventing the spread of fire. If the environmental conditions have high requirements for environmental protection, environmentally friendly flame retardants should be selected. This type of flame retardant does not produce toxic and harmful gases during combustion and is easy to degrade and recycle. For example, certain bio-based flame retardants or inorganic flame retardants (such as zinc borate, aluminum hydroxide, etc.).

[0047] It should also be noted that the required spraying volume is calculated based on the fire level and the area of ​​the spraying area, and precise control is achieved by controlling the spraying speed and spraying time of the spraying device.

[0048] Step 4: Control the UAV to execute a corresponding flame retardant spraying strategy in the flame retardant spraying area.

[0049] During the spraying process, the spraying strategy, including the spraying area, flame retardant type, spraying path and spraying amount, is dynamically adjusted based on real-time feedback of changes in fire intensity and environmental conditions.

[0050] Example 2 like Figure 2 As shown, an embodiment of the present invention provides a control system suitable for spraying flame retardants on a drone, the system comprising: An image recognition module is used to obtain an infrared image and an optical image of the fire scene, fuse the infrared image and the optical image to obtain a fused image, and recognize the fire area image from the fused image; A preliminary analysis module is used to perform spectrum analysis on the image of the fire area and determine the fire source and fire intensity according to the spectrum analysis results; A strategy determination module is used to determine the initial flame retardant spraying area according to the fire area and the fire source point, divide the fire area and the flame retardant spraying area into a number of fire sub-areas and a number of flame retardant spraying areas with the fire source point as the center, and correct the fire intensity of the fire sub-area by considering the terrain factor, vegetation factor and wind factor of the fire sub-area to obtain the regional fire intensity, and correct the initial flame retardant spraying area according to the regional fire intensity to obtain the flame retardant spraying area, and generate a flame retardant spraying strategy; The strategy execution module controls the UAV to execute the corresponding flame retardant spraying strategy in the flame retardant spraying area. It should be understood that a control system suitable for spraying flame retardants on drones provided in an embodiment of the present invention and a control method suitable for spraying flame retardants on drones provided in the above embodiment are based on the same inventive concept. For more specific working principles of each module in the embodiment of the present invention, reference can be made to the above embodiment and will not be repeated in the embodiment of the present invention.

[0051] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. These modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and specification of the present invention.

Claims

1. A control method for spraying flame retardants on drones, characterized in that: include: Acquire an infrared image and an optical image of the fire scene, fuse the infrared image and the optical image to obtain a fused image, and identify a fire area image from the fused image; Performing spectrum analysis on the image of the fire area, and determining the fire source point according to the spectrum analysis result; Determine an initial flame retardant spraying area according to the fire area, modify the initial flame retardant spraying area by taking into account the location of the fire source, wind factor, terrain factor and vegetation factor, obtain the flame retardant spraying area, and generate a flame retardant spraying strategy; The UAV is controlled to execute a corresponding flame retardant spraying strategy in the flame retardant spraying area.

2. A control method for spraying flame retardants on drones according to claim 1, characterized in that: Acquiring an infrared image and an optical image of the fire scene, and fusing the infrared image and the optical image to obtain a fused image includes: Preprocessing the infrared image and the optical image; Using a first coarse feature extraction module and a first attention module to extract features from the preprocessed optical image to obtain a first feature map; Using a second coarse feature extraction module and a second attention module to extract features from the preprocessed optical image to obtain a second feature map; The first feature map and the second feature map are fused to obtain a fused image.

3. A control method for spraying flame retardants on drones according to claim 1, characterized in that: Performing spectrum analysis on the fire area image and determining the fire source point according to the spectrum analysis result includes: Extracting a flame region image from the fused image, performing Fourier transform on the flame region image to obtain a frequency domain representation; Calculate the total energy of each pixel in the fused image in the key spectrum band according to the frequency domain representation, calculate the average energy of a reference point far away from the fire area in the key spectrum band, and use the average energy as a benchmark to calculate the energy ratio between the pixel and the reference point; When the energy ratio is greater than a set value, the pixel point is determined to be a fire source point, and the size of the fire is determined based on the difference between all energy ratios and the set value.

4. A control method for spraying flame retardants on drones according to claim 2, characterized in that: The initial flame retardant spraying area is corrected by taking into account the location of the fire source, wind factor, terrain factor and vegetation factor to obtain the flame retardant spraying area, including: Discretize the regional edge of the fire area into a number of edge fire points, and determine the fire source spread amount according to the position information of the edge fire points and the position information of the fire source point, wherein the fire source spread amount includes a spread direction and a spread distance; Calculate the difference between the energy ratio of all pixels on the straight line between the edge fire point and the fire source point and a set value, and determine the initial fire intensity according to the fire source spread and the difference; The initial fire intensity is corrected by taking into account wind factor, terrain factor and vegetation factor to obtain a target fire intensity; Determine edge points of the flame retardant spraying area according to the target fire intensity, connect the edge points of the flame retardant spraying area to obtain a flame retardant spraying edge, and combine the flame retardant spraying edge and the initial flame retardant spraying area to obtain a flame retardant spraying area.

5. A control method for spraying flame retardants on drones according to claim 4, characterized in that: The regional edge of the fire area is discretized into a plurality of edge fire points, including: determining the discreteness of the edge fire points according to the smoothness of line changes at the regional edge; and discretizing the regional edge into a plurality of edge fire points according to the discreteness.

6. A control method for spraying flame retardants on drones according to claim 4, characterized in that: The initial fire intensity is corrected by taking into account wind factor, terrain factor and vegetation factor, and the formula for obtaining the target fire intensity is: in, Indicates the target fire. Indicates the initial fire. represents the comprehensive correction coefficient, represents the wind factor function, represents the terrain factor function, represents the vegetation factor function, Represents the wind speed vector at the fire source, including the magnitude and direction of the wind speed. It indicates the angle between wind direction and the direction of fire spread. The slope indicates the direction of fire spread. It represents the angle between the slope and the direction of fire spread. Indicates the vegetation type, Indicates vegetation density.

7. A control method for spraying flame retardants on drones according to claim 4, characterized in that: The generated flame retardant spraying strategy includes: determining the type of flame retardant to be sprayed according to the target fire intensity, planning the spraying path of the drone using an optimization algorithm, and generating the flame retardant spraying strategy.

8. A control system suitable for spraying flame retardants on drones, characterized in that: The system comprises: An image recognition module is used to obtain an infrared image and an optical image of the fire scene, fuse the infrared image and the optical image to obtain a fused image, and recognize the fire area image from the fused image; A preliminary analysis module is used to perform spectrum analysis on the image of the fire area and determine the fire source and fire intensity according to the spectrum analysis results; A strategy determination module is used to determine the initial flame retardant spraying area according to the fire area and the fire source point, divide the fire area and the flame retardant spraying area into a number of fire sub-areas and a number of flame retardant spraying areas with the fire source point as the center, and correct the fire intensity of the fire sub-area by considering the terrain factor, vegetation factor and wind factor of the fire sub-area to obtain the regional fire intensity, and correct the initial flame retardant spraying area according to the regional fire intensity to obtain the flame retardant spraying area, and generate a flame retardant spraying strategy; The strategy execution module controls the UAV to execute the corresponding flame retardant spraying strategy in the flame retardant spraying area.

Images