Fire-fighting virtual prevention simulation system method oriented to mountainous region and forest region

Abstract

The invention provides a fire-fighting virtual prevention simulation system and method for a mountain forest region, and the method comprises the steps: extracting a terrain elevation change trend according to a mountain numerical model, and calculating a spreading path of a fire behavior under different terrain conditions in combination with vegetation density distribution; a meteorological data real-time updating module is adopted to obtain current meteorological conditions and humidity data, and prediction parameters of the fire spreading speed are dynamically adjusted; if the terrain elevation change exceeds a preset threshold value, calculating an acceleration effect of the fire behavior on the steep slope, and correcting a spreading path prediction result; if the vegetation density distribution shows valley characteristics, the chimney effect of the fire behavior in valleys is analyzed, and the fire development situation prediction map is updated; according to the soil type and the geological structure, combustion characteristics of different areas are judged, and a calculation formula of the fire spreading speed is adjusted; and adopting a machine learning algorithm, fusing the terrain elevation, the vegetation density and the meteorological condition data, and generating a prediction result of the fire spreading path.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a fire prevention virtual simulation system method for mountainous forest areas. Background Art

[0002] Problem background:

[0003] The core technical challenge facing the forest fire prevention virtualization simulation system when processing high-precision mountain numerical models is how to accurately simulate the spread of fire in complex terrain. The system requires integrating multi-source heterogeneous data, including high-precision remote sensing imagery, drone-based mapping data, and information on vegetation density, soil type, and geological structure collected by ground-based measurement stations. When constructing the mountain numerical model, the spatial resolution and attribute values of these data must be accurately matched to avoid model distortion. The physical mechanisms of fire spread are particularly complex and variable on steep slopes, narrow valleys, and areas with significant differences in vegetation density. The system must accurately calculate the acceleration of fire on steep slopes, evaluate the dynamics of thermal airflow in valleys caused by the chimney effect, and analyze the impact of different vegetation types on burning rates. However, due to the spatial heterogeneity of terrain and vegetation, the model parameters often have nonlinear relationships, increasing the uncertainty of the prediction results. Furthermore, the dynamic changes in real-time meteorological data have a significant impact on the prediction of fire spread paths. The system must process this massive amount of data and update parameters in a short period of time to ensure timely predictions. In this process, how to balance computational accuracy and efficiency and ensure the robustness of the model in complex environments has become a core technical problem that needs to be solved urgently. Summary of the Invention

[0004] The present invention provides a fire prevention virtual simulation system method for mountainous forest areas, which mainly includes:

[0005] Acquire high-precision remote sensing data, drone mapping information, and ground measurement station data to construct a multidimensional mountain numerical model that includes terrain elevation, vegetation density, soil type, and geological structure. Based on the mountain numerical model, extract the trend of terrain elevation changes and, combined with vegetation density distribution, calculate the fire spread path under different terrain conditions. Use a real-time meteorological data update module to obtain current meteorological conditions and humidity data, and dynamically adjust the prediction parameters of the fire spread rate. If the terrain elevation change exceeds the preset threshold, calculate the acceleration effect of the fire on the steep slope and correct the spread path prediction results. If the vegetation density distribution shows valley characteristics, analyze the chimney effect of the fire in the valley and update the fire development situation prediction map. According to the soil type and geological structure, judge the combustion characteristics of different areas and adjust the calculation formula of the fire spread rate. Use a machine learning algorithm to integrate terrain elevation, vegetation density, and meteorological conditions data to generate a prediction result of the fire spread path.

[0006] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0007] The present invention discloses a method for predicting the spread path of mountain forest fires. By constructing a multidimensional mountain numerical model and combining it with real-time meteorological data, the method predicts the fire spread path under complex terrain conditions. The present invention first obtains high-precision remote sensing data and other information, constructs a mountain model that includes factors such as terrain elevation and vegetation density, and then extracts terrain change trends to calculate the fire spread path. At the same time, the present invention adopts a real-time update module for meteorological data to dynamically adjust the prediction parameters, and analyzes the fire acceleration effect and chimney effect based on characteristics such as terrain slope and vegetation distribution to update the fire situation prediction. Finally, the present invention integrates multi-source data and uses a machine learning algorithm to generate fire spread path prediction results, providing a scientific basis for the prevention and control of mountain forest fires. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 The present invention is a flow chart of a fire prevention virtual simulation system method for mountainous forest areas.

[0009] Figure 2 This is a schematic diagram of a fire prevention virtual simulation system method for mountainous forest areas according to the present invention.

[0010] Figure 3 This is another schematic diagram of a fire prevention virtual simulation system method for mountainous forest areas according to the present invention. DETAILED DESCRIPTION

[0011] To further understand the content of the present invention, the present invention is described in detail with reference to the accompanying drawings and examples. The present application is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are intended only to illustrate the relevant invention and are not intended to limit the invention. It should also be noted that, for ease of description, only portions relevant to the invention are shown in the accompanying drawings.

[0012] like Figure 1-3 In this embodiment, a fire prevention virtual simulation system method for mountainous forest areas may specifically include:

[0013] Step S101: Acquire high-precision remote sensing data, UAV mapping information, and ground measurement station data to construct a multi-dimensional mountain numerical model that includes terrain elevation, vegetation density, soil type, and geological structure.

[0014] High-precision information is obtained through remote sensing data and drone mapping. Terrain elevation and vegetation density data are extracted and integrated with ground station information to determine soil type and geological structure characteristics, resulting in a preliminary multidimensional dataset. Data fusion techniques are used to process the preliminary multidimensional dataset, integrating terrain elevation, vegetation density, soil type, and geological structure to generate a standardized mountain feature dataset. The standardized mountain feature dataset is classified using a random forest algorithm to determine the spatial distribution relationship between terrain elevation and vegetation density, generating classified feature distribution data. Based on this classified feature distribution data and combined with geological structure information, an interpolation method is used to generate continuous terrain elevation and vegetation density surfaces, resulting in smooth multidimensional surface data. If the multidimensional surface data deviates from the ground station information by exceeding a preset threshold, interpolation parameters are adjusted to optimize the matching of geological structure and soil type, resulting in a revised multidimensional surface data. Voxelization techniques are used to convert this revised multidimensional surface data into a numerical model, integrating the multidimensional representation of mountain features to produce a multidimensional numerical model of the mountain. For the multidimensional numerical model of mountains, the support vector machine algorithm is used to analyze the correlation between terrain elevation, vegetation density and geological structure, determine the spatial consistency of the model, and obtain the final verification results.

[0015] Specifically, high-precision remote sensing data can be obtained using domestically produced Gaofen-7 satellite imagery, which boasts a panchromatic resolution of 0.8 meters and a multispectral resolution of 3.2 meters. Geometric correction is performed using the RPC model, and image registration is achieved using the SIFT feature matching algorithm, with a mean square error (MSE) of less than 1.5 pixels. Drone mapping uses a DJI M300RTK equipped with an L1 lidar. At an altitude of 150 meters, the point cloud density reaches 200 points per square meter. The point cloud is aligned to ground control points using the ICP iterative closest point algorithm, achieving an elevation accuracy of 5 centimeters. Traverse surveys are performed at the ground survey station using a Trimble S7 total station, with coordinates calculated using the least squares method, and closure errors controlled to within 1 / 20,000. Terrain elevation modeling uses the TIN irregular triangulation algorithm, with a grid spacing of 0.5 meters. Data gaps are filled using kriging interpolation, resulting in a DEM with an elevation MSE of less than 0.3 meters. Vegetation density analysis used the Normalized Difference Vegetation Index (NDVI), calculated as (Near Infrared Band - Red Band) / (Near Infrared Band + Red Band), with a threshold of 0.3-0.8 to distinguish different density levels. Soil type classification was performed using a random forest algorithm, with input parameters including 12 features such as pH (6.2-7.8) and organic matter content (1.5%-4.2%), achieving a classification accuracy of 85%. Geological structure identification used the Canny edge detection algorithm, with a high-low threshold ratio of 1:3. Fault zone orientation was extracted using a Hough transform, and inclination was calculated using the least squares method to fit the plane equation, with an error of less than 3 degrees. Finally, multi-source data was fused using a multi-layer perceptron neural network with three hidden layers and 256 nodes, using the ReLU activation function, to output a 10-meter resolution 3D geological model.

[0016] Step S102 : Based on the mountain numerical model, the terrain elevation change trend is extracted, and combined with the vegetation density distribution, the fire spread path under different terrain conditions is calculated.

[0017] Using terrain elevation data, gradient analysis methods are used to calculate terrain elevation trends, generating elevation change distribution data. This elevation change distribution data is then integrated with vegetation density distribution and weighted overlay methods are used to generate comprehensive terrain-vegetation feature data. Based on this comprehensive terrain-vegetation feature data, spatial gridding techniques are used to construct an initial grid model of fire spread, generating gridded terrain environmental data. Using this gridded terrain environmental data and environmental factors, a diffusion simulation method is used to calculate the fire spread velocity within each grid cell, generating fire velocity distribution data. If the fire velocity distribution data does not match a preset threshold, the diffusion parameters are adjusted to optimize the fire velocity distribution, generating revised velocity distribution data. Based on this revised velocity distribution data, a path planning algorithm is used to calculate the optimal fire spread path under different terrain conditions, determining the final fire path data. Using this final fire path data, volumetric display technology is used to integrate the fire path with terrain elevation and vegetation density to generate a three-dimensional fire spread model and visualize the spread distribution data.

[0018] Based on the matching of fire velocity distribution data with the preset threshold, the diffusion parameter adjustment method is used to obtain the corrected velocity distribution data. Through stereoscopic display technology, the fire path, terrain elevation and vegetation density are integrated to obtain a three-dimensional fire spread model and determine the visualized spread distribution data.

[0019] By comparing the fire speed with a preset threshold, a logical judgment method is used. If the speed exceeds the threshold, the diffusion parameters are adjusted to obtain the corrected distribution data. Based on the corrected distribution data, the fire path information is integrated, and spatial mapping technology is used to generate the initial path grid to obtain terrain-related path data. By superimposing the path data with the terrain elevation, an interpolation calculation method is used to determine the path adjustment data under the influence of elevation. Based on the adjusted path data, the vegetation density distribution is integrated, and weighted analysis technology is used to obtain fire distribution data with comprehensive environmental influences. Based on the fire distribution data with comprehensive environmental influences, stereoscopic display technology is used to generate three-dimensional model data containing terrain and vegetation. Based on the three-dimensional model data, grid division technology is used to obtain the spread distribution data within the spatial unit and determine the final fire spread trend. Data smoothing technology is used to obtain a visual fire spread distribution from the spread distribution data.

[0020] Step S103: Using a real-time meteorological data update module, current meteorological conditions and humidity data are acquired to dynamically adjust the prediction parameters of the fire spread speed.

[0021] Real-time meteorological conditions and humidity data are acquired, and the current environmental status is updated through the data acquisition module to obtain the latest environmental distribution data. Using the latest environmental distribution data, a dynamic adjustment method is used to update the fire spread speed prediction parameters to obtain adjusted speed prediction data. If the adjusted speed prediction data exceeds the preset threshold, the fire spread speed is recalculated using a diffusion simulation method to obtain corrected speed distribution data. Based on the corrected speed distribution data, a spatial grid model of fire spread is constructed using gridding technology to obtain gridded environmental characteristic data. Using the gridded environmental characteristic data, combined with meteorological conditions and humidity data, the direction of fire spread in different grid cells is calculated to determine the fire path distribution data. The fire path distribution data is acquired, and a path planning algorithm is used to optimize the fire spread path to obtain the final path prediction data. The final path prediction data is integrated with real-time updated environmental factors to generate dynamically adjusted fire spread distribution data.

[0022] Specifically, through the real-time update module of meteorological data, the system obtains the current temperature, wind speed, wind direction and humidity data from the weather station every 5 minutes.

[0023] For example, the current temperature is 25 degrees Celsius, the wind speed is 3 meters per second from the southeast, and the humidity is 60%. This data is input into the fire spread prediction model, which uses the Rothermel algorithm for calculations. The Rothermel algorithm dynamically adjusts the fire spread rate based on fuel type, meteorological conditions, and topographic factors. Assuming the current fuel type is dry grass, the model calculates a fire spread rate of 0.5 meters per minute. Humidity data has a significant impact on the fire spread rate. When humidity increases, the moisture content of the fuel increases, slowing the fire spread.

[0024] For example, if humidity increases from 60% to 70%, the model recalculates the fire's spread, assuming a new rate of 0.4 meters per minute. The system also adjusts the direction and extent of the fire's spread based on wind direction and speed.

[0025] For example, southeasterly winds can push a fire northwest, and increased wind speeds can accelerate the spread of a fire. By updating meteorological data in real time, the system can dynamically adjust prediction parameters, providing more accurate fire spread predictions and a scientific basis for emergency response.

[0026] In step S104, if the terrain elevation change exceeds a preset threshold, the acceleration effect of the fire on the steep slope is calculated and the spread path prediction result is corrected.

[0027] If the change in terrain elevation exceeds a preset threshold, the acceleration effect of the fire on the steep slope is calculated using a slope analysis method to obtain fire acceleration data. Using this fire acceleration data and the terrain elevation distribution, an interpolation method is used to generate spatial distribution data for the steep slope effect. Based on the spatial distribution data for the steep slope effect, the acceleration trend of the fire in different slope areas is determined, and the distribution of the acceleration effect is determined. Using the acceleration effect distribution and the initial prediction data of the spread path, a path correction algorithm is used to update the fire spread path to obtain corrected path data. If the deviation between the corrected path data and the initial prediction data exceeds a preset threshold, the spatial distribution of the fire spread is reconstructed using gridding technology to obtain gridded path data. Based on the gridded path data, the elevation change and slope analysis results are integrated to calculate the direction of fire spread in each grid cell and obtain the final path prediction data. The final path prediction data, combined with the acceleration effect distribution, generates dynamically adjusted fire spread distribution data.

[0028] Based on the spatial distribution data of steep slope effects, a trend analysis method is used to obtain the acceleration change characteristics of fire in different slope areas. The distribution of acceleration effects is determined through threshold judgment, and the initial prediction data is updated in combination with a path correction algorithm. If the deviation exceeds the preset threshold, the spatial distribution of fire spread is reconstructed using grid division technology to obtain gridded path data. The elevation change and slope analysis results are integrated to calculate the fire propagation direction of each grid cell to obtain the final path prediction data.

[0029] Using spatial distribution data, trend analysis methods are used to extract the accelerated change characteristics of slope areas and obtain the accelerated change distribution. Based on the accelerated change distribution, combined with threshold judgment, impact distribution data is generated to determine the key areas where fire acceleration occurs. If the deviation between the impact distribution data and the initial prediction exceeds a preset threshold, gridded path data is generated using grid division technology. For the gridded path data, elevation change information is integrated, and the propagation direction data of each grid cell is calculated to obtain the directional distribution. Based on the directional distribution data, the path data is adjusted using interpolation methods to generate a revised path distribution. Based on the revised path distribution, combined with the slope area characteristics, the prediction data is updated to obtain the final prediction result. Based on the final prediction result, visualization technology is used to generate the spatial distribution data of the fire.

[0030] Step S105: If the vegetation density distribution shows valley characteristics, the chimney effect of the fire in the valley is analyzed and the fire development situation prediction map is updated.

[0031] If the vegetation density distribution exhibits valley characteristics, terrain analysis is used to obtain valley boundary data and identify potential areas for the chimney effect. Based on this potential area and combined with wind direction data, the chimney effect intensity data is calculated to determine the effect analysis results. Based on the effect analysis results, interpolation methods are used to generate spatial distribution data for the chimney effect and obtain distribution characteristic information. If the distribution characteristic information indicates a significant effect, the vegetation density data is integrated, the spatial distribution data is updated, and the trend of situation change is determined. Based on this trend of situation change and combined with the initial fire situation data, a path correction algorithm is used to update the predicted data and obtain an adjusted fire situation. The adjusted fire situation is then integrated with the wind speed distribution to calculate the direction of fire propagation in the valley and obtain the final predicted data. Based on the final predicted data and combined with the spatial distribution of the chimney effect, dynamically adjusted fire development situation data is generated.

[0032] Specifically, when analyzing the fire chimney effect where the vegetation density distribution exhibits valley characteristics, we first extract the vegetation density distribution in the valley area using remote sensing imagery and lidar data. For example, we use the NDVI index to divide vegetation coverage levels, setting a threshold of 0.3 or less for low-density areas, 0.3 to 0.6 for medium-density areas, and 0.6 or above for high-density areas. A fluid mechanics model is used to simulate airflow movement under valley terrain, and a CFD algorithm is used to calculate wind speed distribution. Assuming a valley slope of 15 degrees, the wind speed on the windward side can reach 5m / s, while the wind speed drops to 1m / s in the eddy zone formed on the leeward side. Combined with the Rothermel fire spread model, parameters such as a vegetation moisture content of 12% and a combustible load of 2kg / m² are input to calculate the fire line spread rate. The fire intensity in high-density areas can reach 5000kW / m, and in low-density areas it can reach 2000kW / m. The chimney effect was simulated using FDS fire dynamics software, with geometric parameters set to a valley width of 100m and a height difference of 50m. The results showed that high-temperature smoke accumulated at the valley floor, forming a vertical updraft at speeds of up to 8m / s, accelerating the upward spread of the fire. Finally, a cellular automaton was used to update the fire situation prediction map. Each grid cell was 30m x 30m, and the probability of fire spread was iteratively calculated. When adjacent cells had continuous combustibles, the probability of spread increased to 0.8, and when they were discrete, it decreased to 0.2. A thermal map was dynamically generated, showing fire intensity, spread direction, and speed.

[0033] Step S106: Determine the combustion characteristics of different areas based on soil type and geological structure, and adjust the calculation formula for the fire spread rate.

[0034] Soil type and geological structure data are used to obtain combustion characteristic distribution information and determine the initial combustion characteristic values for each region. Based on this combustion characteristic distribution information and combined with terrain boundary data, regional distribution characteristics are determined to obtain combustion characteristic data for the divided regions. If the combustion characteristic data for the divided regions indicates a high organic matter content, the initial parameters are adjusted using a preset threshold to obtain corrected parameter data. Using this corrected parameter data, wind direction distribution information is integrated to determine the spread speed adjustment coefficient and obtain regional spread speed distribution data. Using this regional spread speed distribution data and combined with geological structure characteristics, the direction of fire spread is determined to obtain direction-adjusted speed data. Based on this direction-adjusted speed data, interpolation methods are used to generate spatial distribution information for the spread speed to obtain the final adjusted fire spread speed data. Using this final adjusted fire spread speed data, combined with soil type distribution, changes in combustion intensity in each region are determined to obtain dynamically adjusted fire intensity data.

[0035] Step S107: Using a machine learning algorithm, the terrain elevation, vegetation density, and meteorological condition data are integrated to generate a prediction result of the fire spread path.

[0036] A machine learning algorithm fuses terrain elevation, vegetation density, and meteorological condition data to generate initial predicted distribution data for the fire spread path. Using a random forest algorithm, the terrain elevation and vegetation density data are used to determine the characteristic values of elevation change and density distribution, generating feature-adjusted distribution data. This feature-adjusted distribution data is then integrated with meteorological condition data to determine the weight coefficients of the influence of these conditions, generating weight-adjusted path prediction data. Based on this weight-adjusted path prediction data, combined with elevation change and density distribution, the spatial distribution characteristics of the fire spread are determined, generating spatially adjusted spread data. If the spatially adjusted spread data indicates areas of high density, the directional coefficient of the path prediction is adjusted using a preset threshold to generate direction-corrected path data. Interpolation methods are used to generate spatially continuous distribution information for the fire spread path from the direction-corrected path data, generating the final predicted path distribution data. The final predicted path distribution data is then integrated with the influence of the conditions and distribution generation information to determine the dynamic trends of the fire spread and generate dynamically adjusted prediction results.

[0037] Specifically, terrain data was first acquired through a digital elevation model. SRTM elevation data with a resolution of 30 meters was used, and the bilinear interpolation algorithm was used to increase the resolution to 10 meters. Topographic factors such as slope and aspect were calculated, with the slope threshold set at 15 degrees or above, marking areas as high-risk. Vegetation density data was derived from the NDVI index of Landsat 8. Areas with NDVI values greater than 0.6 were classified as high-density vegetation areas using a support vector machine algorithm. Combined with historical fire data, statistics show that fire spread rates in such areas can reach 0.8 meters per second. Meteorological data was connected to real-time wind speed and direction observation station data. When the wind speed exceeded 5 meters per second, a random forest algorithm was used to predict the dominant direction of the fire. Feature importance analysis showed that wind speed contributed 42% to the prediction results. The three types of data were fed into an improved cellular automation model, using 30×30 meter cells. Each cell contained 12-dimensional features, including elevation, vegetation type, and humidity. After extracting spatial features using a convolutional neural network, an LSTM network was used to predict fire spread over the next six hours. The inclusion of an attention mechanism during training improved the model's prediction accuracy to 89% under hot and dry conditions. The final output was generated using a Gaussian mixture model to generate a probabilistic heat map. Areas with a probability of fire spread greater than 70% were marked with a red warning. The system automatically updated the predictions every 15 minutes and overlaid them with satellite infrared imagery for cross-validation.

[0038] Based on the changes in terrain elevation and vegetation density distribution, the random forest algorithm is used to integrate meteorological condition data to determine the spatial distribution characteristics of the fire spread path and obtain dynamically adjusted fire spread prediction result data.

[0039] Using a random forest algorithm to fuse terrain elevation and vegetation density data, we obtain characteristic information about elevation changes and density distribution, generating preliminary spatial distribution data. This preliminary spatial distribution data is then integrated with meteorological condition data to determine the weight coefficients of these conditions, generating weight-adjusted distribution data. Using this weight-adjusted distribution data, we determine the spatial distribution characteristics of fire spread and obtain characteristic-corrected path data. If the high-density distribution area in the characteristic-corrected path data exceeds a preset threshold, the path's directional coefficient is adjusted to obtain direction-adjusted path data. Interpolation is then used to process this direction-adjusted path data, generating continuous distribution information about the fire spread path and generating smoothed distribution data. This smoothed distribution data is then integrated with meteorological condition trends to determine the dynamic characteristics of fire spread and generate dynamically adjusted prediction data. Based on this dynamically adjusted prediction data, combined with the characteristics of elevation changes and density distribution, we obtain the final fire spread path distribution and obtain comprehensively adjusted result data.

[0040] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to specific embodiments. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A fire prevention virtual simulation system method for mountainous forest areas, characterized by: The method comprises: Acquire high-precision remote sensing data, drone mapping information, and ground measurement station data to construct a multidimensional mountain numerical model that includes terrain elevation, vegetation density, soil type, and geological structure. Based on the mountain numerical model, extract the trend of terrain elevation changes and, combined with vegetation density distribution, calculate the fire spread path under different terrain conditions. Use a real-time meteorological data update module to obtain current meteorological conditions and humidity data, and dynamically adjust the prediction parameters of the fire spread rate. If the terrain elevation change exceeds the preset threshold, calculate the acceleration effect of the fire on the steep slope and correct the spread path prediction results. If the vegetation density distribution shows valley characteristics, analyze the chimney effect of the fire in the valley and update the fire development situation prediction map. According to the soil type and geological structure, judge the combustion characteristics of different areas and adjust the calculation formula of the fire spread rate. Use a machine learning algorithm to integrate terrain elevation, vegetation density, and meteorological conditions data to generate a prediction result of the fire spread path.

2. The method according to claim 1, characterized in that The acquisition of high-precision remote sensing data, UAV mapping information, and ground survey station data to construct a multi-dimensional mountain numerical model that includes terrain elevation, vegetation density, soil type, and geological structure includes: High-precision information is obtained through remote sensing data and drone mapping, terrain elevation and vegetation density data are extracted, and soil type and geological structure characteristics are determined by integrating ground station information to obtain a preliminary multidimensional data set; Data fusion techniques were used to process the preliminary multidimensional dataset, integrating terrain elevation, vegetation density, soil type, and geological structure to generate a standardized mountain feature dataset; The standardized mountain feature dataset was classified using the random forest algorithm to determine the spatial distribution relationship between terrain elevation and vegetation density, and to obtain the classified feature distribution data. Based on the classified feature distribution data and combined with geological structure information, the interpolation method is used to generate continuous terrain elevation and vegetation density surfaces to obtain smooth multi-dimensional surface data; If the deviation between the multidimensional surface data and the ground station information exceeds a preset threshold, the interpolation parameters are adjusted to optimize the matching degree of geological structure and soil type to obtain the corrected multidimensional surface data; The modified multidimensional surface data is converted into a numerical model through voxelization technology, and the multidimensional construction of mountain features is integrated to obtain a multidimensional numerical model of the mountain. For the multidimensional numerical model of mountainous areas, the support vector machine algorithm is used to analyze the correlation between terrain elevation, vegetation density and geological structure, determine the spatial consistency of the model, and obtain the final verification results.

3. The method according to claim 1, characterized in that The above method extracts the terrain elevation change trend based on the mountain numerical model and combines it with the vegetation density distribution to calculate the fire spread path under different terrain conditions, including: Using terrain elevation data and gradient analysis methods, we can calculate the changing trend of terrain elevation and obtain elevation change distribution data. Based on the elevation change distribution data, the vegetation density distribution is integrated and the weighted overlay method is used to generate terrain-vegetation comprehensive characteristic data; Based on the terrain-vegetation comprehensive characteristic data, spatial grid division technology is used to construct an initial grid model of fire spread and obtain gridded terrain environment data; By combining gridded terrain environment data with environmental factors and adopting diffusion simulation method, the fire spread speed in different grid cells is calculated to obtain the fire speed distribution data. If the fire velocity distribution data does not match the preset threshold, the fire velocity distribution is optimized by adjusting the diffusion parameters to obtain the corrected velocity distribution data; Based on the corrected velocity distribution data, a path planning algorithm is used to calculate the optimal fire spread path under different terrain conditions and determine the final fire path data; The final fire path data is used, and volumetric display technology is used to fuse the fire path with the terrain elevation and vegetation density to generate a three-dimensional fire spread model and obtain visualized spread distribution data; it also includes: according to the matching of the fire velocity distribution data with the preset threshold, a diffusion parameter adjustment method is used to obtain corrected velocity distribution data, and the fire path, terrain elevation and vegetation density are fused through volumetric display technology to obtain a three-dimensional fire spread model and determine the visualized spread distribution data.

4. The method according to claim 1, wherein The real-time meteorological data update module is used to obtain current meteorological conditions and humidity data and dynamically adjust the fire spread speed prediction parameters, including: Obtain real-time meteorological conditions and humidity data, update the current environmental status through the data acquisition module, and obtain the latest environmental distribution data; Using the latest environmental distribution data, a dynamic adjustment method is used to update the fire spread speed prediction parameters to obtain adjusted speed prediction data; If the adjusted velocity prediction data exceeds the preset threshold, the fire spread velocity is recalculated using the diffusion simulation method to obtain the corrected velocity distribution data; According to the corrected velocity distribution data, a spatial grid model of fire spread is constructed using grid division technology to obtain gridded environmental characteristic data. By combining gridded environmental characteristic data with meteorological conditions and humidity data, the fire propagation direction in different grid cells is calculated to determine the fire path distribution data. Obtain fire path distribution data, use path planning algorithms to optimize the fire spread path, and obtain the final path prediction data; The final path prediction data is integrated with real-time updated environmental factors to generate dynamically adjusted fire spread distribution data.

5. The method according to claim 1, wherein If the terrain elevation change exceeds a preset threshold, the acceleration effect of the fire on the steep slope is calculated and the spread path prediction result is corrected, including: If the terrain elevation change exceeds a preset threshold, the acceleration effect of the fire on the steep slope is calculated using the slope analysis method to obtain fire acceleration data; The spatial distribution data of steep slope impact is generated by interpolation method through fire acceleration data combined with terrain elevation distribution; Based on the spatial distribution data of steep slope effects, the acceleration trend of fire in different slope areas is obtained to determine the distribution of acceleration effects; By accelerating the impact distribution and combining the initial prediction data of the spread path, a path correction algorithm is used to update the fire spread path and obtain the corrected path data; If the deviation between the corrected path data and the initial prediction data exceeds a preset threshold, the spatial distribution of the fire spread is reconstructed using gridding technology to obtain gridded path data; Based on the gridded path data, the elevation change and slope analysis results are integrated to calculate the direction of fire spread in each grid cell and obtain the final path prediction data; The final path prediction data is combined with the acceleration impact distribution to generate dynamically adjusted fire spread distribution data. This also includes: based on the spatial distribution data of steep slope impact, a trend analysis method is used to obtain the acceleration change characteristics of the fire in different slope areas, the acceleration impact distribution is determined through threshold judgment, and the initial prediction data is updated in combination with the path correction algorithm. If the deviation exceeds the preset threshold, the spatial distribution of the fire spread is reconstructed using grid division technology to obtain gridded path data. The elevation change and slope analysis results are integrated to calculate the fire propagation direction of each grid unit to obtain the final path prediction data.

6. The method according to claim 1, characterized in that If the vegetation density distribution shows valley characteristics, the chimney effect of the fire in the valley is analyzed and the fire development situation prediction map is updated, including: If the vegetation density distribution shows valley characteristics, the boundary data of the valley characteristics can be obtained through terrain analysis to obtain the potential area of chimney effect; According to the potential area, combined with the wind direction data, the intensity data of the chimney effect is calculated to determine the effect analysis results; Based on the effect analysis results, the spatial distribution data of the chimney effect is generated using the interpolation method to obtain the distribution characteristic information; If the distribution characteristic information shows a significant effect, the vegetation density data is fused and the spatial distribution data is updated to obtain the trend of situation change; According to the trend of fire situation changes and combined with the initial fire situation data, the path correction algorithm is used to update the predicted data to obtain the adjusted fire situation; The final forecast data is obtained by combining the adjusted fire situation with the wind speed distribution and calculating the direction of fire spread in the valley. Based on the final prediction data and combined with the spatial distribution of the chimney effect, dynamically adjusted fire development situation data are generated.

7. The method according to claim 1, characterized in that The calculation formula for fire spread rate is adjusted based on soil type and geological structure to determine the combustion characteristics of different areas, including: Obtain combustion characteristic distribution information through soil type and geological structure data and determine the initial combustion characteristic value of each area; According to the combustion characteristic distribution information and combined with the terrain boundary data, the regional distribution characteristics are judged to obtain the combustion characteristic data of the divided region; If the combustion characteristic data of the divided area shows a high organic matter content, the initial parameters are adjusted by a preset threshold to obtain corrected parameter data; The modified parameter data is used to integrate the wind direction distribution information, determine the adjustment coefficient of the spread speed, and obtain the regional spread speed distribution data; By combining regional spread velocity distribution data with geological structural characteristics, the direction of fire spread is determined and velocity data after direction adjustment is obtained; According to the velocity data after direction adjustment, the spatial distribution information of the spread velocity is generated by using the interpolation method to obtain the final adjusted fire spread velocity data; By integrating the final adjusted fire spread rate data with the soil type distribution, the changes in burning intensity in each area are judged to obtain dynamically adjusted fire data.

8. The method according to claim 1, characterized in that The machine learning algorithm is used to integrate terrain elevation, vegetation density, and meteorological conditions data to generate predictions of fire spread paths, including: Using machine learning algorithms to fuse terrain elevation, vegetation density, and meteorological condition data, we generate initial predicted distribution data for fire spread paths. The random forest algorithm is used to determine the characteristic values of elevation change and density distribution based on terrain elevation and vegetation density data, and the distribution data after feature adjustment is obtained; By integrating the distribution data after feature adjustment with the meteorological condition data, the weight coefficient of the conditional influence is determined to obtain the weight-adjusted path prediction data; Based on the weighted path prediction data, combined with elevation changes and density distribution, the spatial distribution characteristics of fire spread are determined to obtain spatially adjusted spread data; If the spatially adjusted spread data shows a high-density distribution area, the direction coefficient of the path prediction is adjusted by a preset threshold to obtain the direction-corrected path data; The interpolation method is used to generate the spatial continuous distribution information of the fire spread path through the direction-corrected path data, and the final predicted path distribution data is obtained; By using the final predicted path distribution data, integrating the conditional influence and distribution generation information, the dynamic change trend of the fire spread is judged, and the dynamically adjusted prediction result data is obtained; it also includes: according to the terrain elevation changes and vegetation density distribution, the random forest algorithm is used to integrate the meteorological condition data to judge the spatial distribution characteristics of the fire spread path and obtain the dynamically adjusted fire spread prediction result data.

Images