Forest fire spreading simulation method and system based on unmanned aerial vehicle LiDAR technology

Abstract

The invention provides a forest fire spreading simulation method and system based on an unmanned aerial vehicle LiDAR technology, and the method comprises the steps: collecting forest region data based on the unmanned aerial vehicle LiDAR technology, a camera and a sensor; wherein the forest area data comprises LiDAR point cloud data, thermal infrared data, image data and meteorological data; based on the forest region data, performing forest region feature extraction to obtain feature parameters; based on the characteristic parameters, constructing a fire spread dynamic model through a fire spread dynamics reinforcement learning framework and a vertical-horizontal fire spread coupling algorithm; based on the fire spreading dynamic model, fire spreading simulation and prediction are carried out, dynamic early warning is carried out based on simulation and prediction results, and a fire fighting and fire prevention decision is obtained. The technical scheme of the invention has important significance for protecting forest resources and maintaining ecological balance.

Description

Technical Field

[0001] The present invention belongs to the technical field of dynamic simulation of forest fire spread, and particularly relates to a method and system for simulating forest fire spread based on UAV LiDAR technology. Background Art

[0002] With the intensification of global climate change and the frequent occurrence of extreme climate events such as high temperature and drought, the frequency and intensity of forest fires have increased significantly. Traditional forest fire spread models usually rely on ground observations and satellite data, but these methods have significant limitations. The coverage of ground observations is limited, and the data acquisition cycle is long, making it difficult to meet the needs of real-time fire situation simulation; while the resolution of satellite data is usually 10 - 30 meters, and the time resolution is 5 - 16 days, which cannot depict the complex forest fuel structure or rapidly changing fire situation in detail. In addition, the accuracy of existing models in areas with complex terrain and heterogeneous fuel distribution is low, and it is difficult to comprehensively capture the dynamic characteristics of forest fire spread.

[0003] In recent years, the rapid development of UAV LiDAR technology has provided new data support for forest fire spread simulation. LiDAR point cloud data has high spatial resolution (more than 200 points per square meter), high precision, and three-dimensional structure characteristics, and can depict the forest fuel distribution and terrain characteristics in detail. The application of this technology in subtropical forest environments has particularly great potential, which can make up for the deficiencies of traditional methods and provide an important technical basis for improving the accuracy and real-time performance of forest fire spread simulation. Summary of the Invention

[0004] In order to solve the problems existing in the prior art, the present invention uses UAV LiDAR technology and a variety of sensors to collect multi-modal data to achieve dynamic prediction and early warning of forest fire spread trends.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] A method for simulating forest fire spread based on UAV LiDAR technology, the method comprising:

[0007] Collecting forest area data based on UAV LiDAR technology, a camera, and sensors; wherein, the forest area data includes LiDAR point cloud data, thermal infrared data, image data, and meteorological data;

[0008] Extracting forest area features based on the forest area data to obtain feature parameters;

[0009] Constructing a fire spread dynamic model based on the feature parameters through a fire spread dynamics reinforcement learning framework and a vertical-horizontal fire spread coupling algorithm;

[0010] Based on the fire spread dynamic model, fire spread simulation and prediction are carried out, and dynamic early warning is carried out based on the simulation and prediction results to obtain fire fighting and prevention decisions.

[0011] Preferably, the characteristic parameters include terrain parameters, fuel parameters and meteorological parameters;

[0012] The terrain parameters include the altitude, slope and aspect of the forest area;

[0013] The fuel parameters include surface fuel type, fuel moisture characteristics, surface fuel load, canopy density, vertical fuel distribution and canopy gap distribution;

[0014] The meteorological parameters include temperature, humidity, wind speed, wind direction, atmospheric pressure and light intensity.

[0015] Preferably, the method for constructing the fire spread dynamic model includes:

[0016] Based on the terrain parameters, fuel parameters and meteorological parameters, a deep reinforcement learning model is constructed, and the surface fire and crown fire spread processes are regarded as Markov decision processes to obtain the dynamic fire line position and intensity;

[0017] Based on the flame heat radiation transfer equation and the air turbulence model, combined with the canopy gap distribution, calculate the flame jump distance and probability to obtain the coupled model of vertical spread and horizontal diffusion of crown fire;

[0018] Based on the coupled model, simulate and obtain the three-dimensional fire spread path;

[0019] Based on the dynamic fire line position and intensity and the three-dimensional fire spread path, construct the fire spread dynamic model.

[0020] Preferably, the method for obtaining the dynamic fire line position and intensity includes:

[0021] Discretize the terrain parameters, the fuel parameters and the meteorological parameters;

[0022] Based on the convolutional neural network, obtain the spatial distribution characteristics of the discretized terrain parameters and discretized fuel parameters, and capture the temporal characteristics of the discretized meteorological parameters based on the long short-term memory network;

[0023] Based on the spatial distribution characteristics and the temporal characteristics, construct a state space; take the fire line position and intensity as the action space;

[0024] Take the state space as the input of the deep Q network, take the action space as the output of the deep Q network, update the network weights by minimizing the loss function, and train the agent;

[0025] The agent selects the optimal action according to the current state, simulates the Markov decision process, and obtains the dynamic fire line position and intensity.

[0026] Preferably, the method for obtaining the coupled model of the vertical spread and horizontal diffusion of the forest canopy fire includes:

[0027] Adopt the radiation heat transfer theory to establish the flame thermal radiation transfer equation, and obtain the propagation and attenuation process of the flame thermal radiation in the forest canopy based on the flame thermal radiation transfer equation;

[0028] Based on the air turbulence intensity, adopt the air turbulence model to simulate the influence process of air turbulence on the flame spread;

[0029] Based on the propagation and attenuation process of the flame thermal radiation in the forest canopy and the influence process of air turbulence on the flame spread, combined with the canopy gap distribution, calculate the flame jump distance and probability, and obtain the coupled model of the vertical spread and horizontal diffusion of the forest canopy fire.

[0030] Preferably, it further includes integrating LiDAR point cloud data, fire spread simulation and prediction results, and real-time meteorological data, establishing a three-dimensional dynamic visualization platform, dynamically displaying the movement trend of the fire line through heat maps and particle effects, and dividing the fire risk levels.

[0031] The present invention also provides a forest fire spread simulation system based on UAV LiDAR technology for implementing the method, including:

[0032] A data acquisition module for collecting forest area data based on UAV LiDAR technology, cameras, and sensors; wherein, the forest area data includes LiDAR point cloud data, thermal infrared data, image data, and meteorological data;

[0033] A feature parameter extraction module for extracting forest area features based on the forest area data to obtain feature parameters;

[0034] A fire spread dynamic model construction module for constructing a fire spread dynamic model based on the feature parameters through a fire spread dynamics reinforcement learning framework and a vertical-horizontal fire spread coupling algorithm;

[0035] A fire situation simulation module for simulating and predicting the fire spread based on the fire spread dynamic model, and performing dynamic early warning based on the simulation and prediction results to obtain fire fighting and fire prevention decisions.

[0036] Preferably, in the feature parameter extraction module, the feature parameters include terrain parameters, fuel parameters, and meteorological parameters;

[0037] The terrain parameters include the altitude, slope, and aspect of the forest area;

[0038] The fuel parameters include surface fuel type, fuel moisture characteristics, surface fuel load, canopy density, vertical fuel distribution, and canopy gap distribution;

[0039] The meteorological parameters include temperature, humidity, wind speed, wind direction, atmospheric pressure, and light intensity.

[0040] Compared with the prior art, the beneficial effects of the present invention are as follows: Through the fusion of multi-source data such as UAV LiDAR technology, the fire points can be identified more accurately and the trend of fire spread can be predicted, thereby improving the accuracy of early warning. Based on the simulation and prediction results of the fire spread dynamic model, more scientific and reasonable fire-fighting and fire-prevention decisions can be formulated, improving the rescue efficiency and safety. Through dynamic early warning and real-time data update, the fire situation can be responded to faster, the rescue response time can be shortened, and the losses caused by the fire can be reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions of the present invention, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0042] Figure 1 It is a schematic flow chart of a method for simulating forest fire spread based on UAV LiDAR technology in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0043] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0044] To make the above objects, features, and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below with reference to the drawings and specific embodiments.

[0045] Embodiment 1

[0046] As Figure 1 shown, a method for simulating forest fire spread based on UAV LiDAR technology, the method includes:

[0047] S1: Collect forest area data based on UAV LiDAR technology, cameras, and sensors. Among them, the forest area data includes LiDAR point cloud data, thermal infrared data, image data, and meteorological data. Specifically, use a thermal infrared camera to collect thermal infrared data and an RGB camera to collect RGB image data.

[0048] In this embodiment, the dynamic flight path of the UAV is optimized based on real-time meteorological conditions (such as wind speed and temperature) and terrain complexity (by pre-collecting DEM data). The three-dimensional forest model is discretized into a grid map, and each grid node stores terrain height, fire intensity, and risk level. An initial path from the starting point to the target point is searched through the A* algorithm. Based on the initial path generated by the A* algorithm, the local path is optimized through dynamic programming to reduce the path length and risk. The UAV automatically adjusts the flight height (30 - 150 meters) and overlap rate (30% - 50%) according to the fire risk level of the fire area to ensure that the data collection density in high-risk areas is greater than or equal to 300 points / m², and the LiDAR point cloud data is compressed and transmitted in real time through edge computing to reduce latency.

[0049] S2: Based on the forest area data, extract the forest area features to obtain feature parameters. A further implementation method is that the feature parameters include terrain parameters, fuel parameters, and meteorological parameters;

[0050] The terrain parameters include the altitude, slope, and aspect of the forest area. In this embodiment, a multi-scale point cloud hierarchical processing method is adopted, and the ground points and non-ground points are separated through a density clustering algorithm. For complex terrains (such as steep slopes and gullies), an adaptive filtering algorithm is introduced to dynamically adjust the filtering window size (3×3 to 15×15 pixels) to eliminate vegetation interference. Combine the Kriging interpolation method to generate a high-precision DEM (resolution ≤ 0.5 meters), and use the vector differential algorithm of slope and aspect to extract terrain parameters. In this embodiment, the barometric sensor carried by the UAV is synchronously calibrated with the LiDAR elevation data to eliminate the influence of atmospheric pressure fluctuations on altitude measurement. Combine the real-time wind speed data to establish a relationship model between terrain undulation and local turbulence intensity, and dynamically correct the influence weight of aspect on the spread of the fire.

[0051] Fuel parameters include surface fuel type, fuel moisture characteristics, surface fuel load, canopy density, vertical fuel distribution, and canopy gap distribution; in this embodiment, the surface fuel type is obtained through image recognition based on RGB image data. The fuel moisture characteristics are obtained using the surface temperature data of an infrared thermal camera, air temperature, and relative humidity. Combining the ground reflection intensity of LiDAR point cloud with the texture characteristics of RGB images, a convolutional neural network (CNN) model is trained to classify and identify surface fuel types such as litter, shrubs, and tree trunks, and the load per unit area is inverted through a regression model. The voxelization method is used to segment the LiDAR point cloud (voxel size 0.1m 3 ), the point cloud density within each voxel layer is counted to generate a vertical profile of the canopy; combining the canopy temperature gradient inverted from thermal infrared data to optimize the vertical flammability assessment of the fuel. Based on a three-dimensional point cloud skeleton extraction algorithm, the geometric characteristics (area, shape index) of the canopy gaps are identified, and their spatial connectivity is calculated.

[0052] Meteorological parameters include temperature, humidity, wind speed, wind direction, atmospheric pressure, and light intensity.

[0053] S3: Based on the characteristic parameters, a wildfire spread dynamic model is constructed through a wildfire spread dynamics reinforcement learning framework and a vertical-horizontal wildfire spread coupling algorithm; a further implementation method lies in that the method for constructing the wildfire spread dynamic model includes:

[0054] S31: Based on the terrain parameters, fuel parameters, and meteorological parameters, a deep reinforcement learning model is constructed, and the surface fire and crown fire spread processes are regarded as Markov decision processes to obtain the dynamic fire line position and intensity; a further implementation method lies in that the method for obtaining the dynamic fire line position and intensity includes:

[0055] The terrain parameters, fuel parameters, and meteorological parameters are discretized; specifically, in this embodiment, the terrain parameters (altitude, slope, aspect) are divided into regular grids (such as 10m×10m), and each grid cell stores the terrain feature values. Bilinear interpolation is used to fill in the missing data to ensure spatial continuity. The terrain complexity index (TCI) is introduced to dynamically adjust the grid resolution, and higher resolution (5m×5m) is used in high-complexity areas (such as steep slopes, gullies). The fuel parameters (surface fuel load, canopy density, vertical fuel distribution, etc.) are discretized in layers, with each layer thickness of 0.5m, and the grid cells are aligned with the terrain parameters. The fuel distribution is smoothed through a three-dimensional convolutional kernel to eliminate noise. The fuel flammability index (FCI) is proposed to comprehensively consider the fuel load, humidity, and vertical distribution to quantify the flammability of the grid cells. The meteorological parameters (temperature, humidity, wind speed, wind direction, etc.) are discretized according to the time series, with a time step of 5 minutes. The local meteorological features are extracted through a sliding window method to reduce data fluctuations.

[0056] Based on the convolutional neural network, obtain the spatial distribution characteristics of discretized terrain parameters and discretized fuel parameters, and capture the temporal characteristics of discretized meteorological parameters based on the long short-term memory network; specifically, for the extraction of spatial distribution characteristics, construct a multi-layer CNN network, with the input being the discretized grid data of terrain and fuel parameters, and the output being a high-dimensional feature vector. The network structure includes:

[0057] Input layer: Multi-channel grid data of terrain + fuel parameters.

[0058] Convolutional layer: 3×3 convolutional kernel, extracting local spatial features. Specifically, introduce the attention mechanism (Attention) to dynamically weight the importance of terrain and fuel features, enhancing the pertinence of feature extraction.

[0059] Pooling layer: Max pooling, reducing the feature dimension.

[0060] Fully connected layer: Output a 128-dimensional feature vector. Concatenate the terrain and fuel feature vectors, and reduce the dimension to 64 through principal component analysis (PCA), which is used as part of the state space.

[0061] Regarding the extraction of temporal characteristics, introduce a bidirectional LSTM (BiLSTM) to capture the meteorological change trends in the past and future simultaneously, and the output is a 64-dimensional temporal feature vector.

[0062] Based on the spatial distribution characteristics and temporal characteristics, construct a state space; use the fire line position and intensity as the action space; specifically, concatenate the spatial features (64-dimensional) and temporal features (64-dimensional) to construct a 128-dimensional state space vector.

[0063] Use the state space as the input of the deep Q network (DQN), and the action space as the output of the deep Q network. Update the network weights by minimizing the loss function to train the agent; specifically, the state space is a 128-dimensional feature vector, representing the comprehensive state of terrain, fuel, and meteorology. The action space is the fire line position (grid coordinates) and intensity (low / medium / high), with a total of N×3 actions (N is the number of grids).

[0064] The DQN network structure includes:

[0065] Input layer: 128-dimensional state space vector.

[0066] Hidden layer: 2 layers of fully connected layers (256 neurons), with ReLU activation function.

[0067] Output layer: N×3-dimensional action space vector, with Softmax activation function.

[0068] Loss function:

[0069] Among them, Q(s,a) is the current Q value, r is the immediate reward, and γ is the discount factor.

[0070] Introduce Prioritized Experience Replay (PER), prioritize training high-error samples to accelerate convergence, that is, store state transition samples (s,a,r,s′) and randomly sample for training. Regularly update the target Q network to stabilize the training process.

[0071] The agent selects the optimal action according to the current state, simulates the Markov decision-making process, and obtains the dynamic fire line position and intensity. Specifically, according to the current state and action, update the fire line position and intensity, and combine with meteorological changes to predict the next state.

[0072] Reward function: r = w1 * fire intensity + w2 * spread speed + w3 * risk area coverage,

[0073] Among them, w1, w2, and w3 are weight coefficients, which are calibrated through experiments.

[0074] S32: Based on the flame heat radiation transfer equation and the air turbulence model, combined with the canopy gap distribution, calculate the flame jump distance and probability, and obtain the coupling model of the vertical spread and horizontal diffusion of the canopy fire;

[0075] A further implementation method is that the method for obtaining the coupling model of the vertical spread and horizontal diffusion of the canopy fire includes:

[0076] S321: Adopt the radiation heat transfer theory to establish the flame heat radiation transfer equation, and obtain the propagation and attenuation process of the flame heat radiation in the canopy based on the flame heat radiation transfer equation; in this embodiment, adopt the non-gray body radiation model, regard the flame as a multi-wavelength radiation source, and combine with the canopy fuel density distribution extracted by LiDAR to calculate the radiation intensity attenuation coefficient (κ λ ) and the scattering coefficient (σ λ ). Solve the RTE by the Discrete Ordinate Method (DOM), and the formula is:

[0077]

[0078] Among them, I λ is the radiation intensity, s is the path length, and Φ is the scattering phase function. By introducing the vertical distribution data of the canopy fuel, dynamically adjust κ λ and σ λ , and improve the scene adaptability of the radiation heat transfer calculation.

[0079] S322: Based on the air turbulence intensity, an air turbulence model is adopted to simulate the influence process of air turbulence on flame spread; in this embodiment, a three-dimensional turbulent wind field model based on large eddy simulation (LES) is established to solve the Navier-Stokes equation and capture the wind speed pulsation and turbulent vortices in the canopy area. The flame is regarded as a heat source term, and the buoyancy effect is embedded in the momentum equation through the Boussinesq approximation to dynamically update the temperature field and velocity field. Among them, regarding the velocity field, the fire spread rate formula is improved based on the Rothermel model, and a turbulent kinetic energy (TKE) correction term is introduced:

[0080]

[0081] where R 0 is the reference spread rate, and α is the turbulence intensity coefficient, which is calibrated through experiments.

[0082] S323: Based on the propagation and attenuation process of flame thermal radiation in the forest canopy and the influence process of air turbulence on flame spread, combined with the canopy gap distribution, calculate the flame jump distance and probability to obtain a coupled model of vertical flame spread and horizontal diffusion in the forest canopy.

[0083] Specifically, considering the preheating effect of flame thermal radiation on combustibles and the promoting effect of air turbulence on flame jumping, a flame jumping mechanism in the forest canopy gap is established, including flame propagation mode, jumping conditions, and other influencing factors, etc. Combining the canopy gap distribution and the flame jumping mechanism, considering the influence of different fire intensities, wind speeds, and combustible types on the flame jump distance and probability, a random walk algorithm is used to simulate the flight trajectory of fire sparks, and a calculation model of flame jump distance and probability is established.

[0084] S33: Based on the coupled model, simulate and obtain the three-dimensional fire spread path;

[0085] S34: Based on the dynamic fire line position and intensity and the three-dimensional fire spread path, construct a dynamic fire spread model.

[0086] S4: Based on the dynamic fire spread model, conduct fire spread simulation and prediction, and based on the simulation and prediction results, conduct dynamic early warning to obtain fire fighting and prevention decisions.

[0087] A further implementation method is that it also includes integrating LiDAR point cloud data, fire spread simulation and prediction results, and real-time meteorological data to establish a three-dimensional dynamic visualization platform, and dynamically display the moving trend of the fire line through heat maps and particle effects, and divide the fire risk levels.

[0088] Embodiment 2

[0089] The present invention also provides a forest fire spread simulation system based on UAV LiDAR technology for implementing the method, including:

[0090] A data acquisition module, configured to collect forest area data based on UAV LiDAR technology, cameras, and sensors; wherein, the forest area data includes LiDAR point cloud data, thermal infrared data, image data, and meteorological data;

[0091] A feature parameter extraction module, configured to extract forest area features based on the forest area data to obtain feature parameters;

[0092] A fire spread dynamic model construction module, configured to construct a fire spread dynamic model based on the feature parameters through a fire spread dynamics reinforcement learning framework and a vertical-horizontal fire spread coupling algorithm;

[0093] A fire situation simulation module, configured to simulate and predict the fire spread based on the fire spread dynamic model, and perform dynamic early warning based on the simulation and prediction results to obtain fire fighting and prevention decisions.

[0094] In a further implementation manner, in the feature parameter extraction module, the feature parameters include terrain parameters, fuel parameters, and meteorological parameters;

[0095] The terrain parameters include the altitude, slope, and aspect of the forest area;

[0096] The fuel parameters include surface fuel type, fuel moisture characteristics, surface fuel load, canopy density, vertical fuel distribution, and canopy gap distribution;

[0097] The meteorological parameters include temperature, humidity, wind speed, wind direction, atmospheric pressure, and light intensity.

[0098] The above-described embodiments are only descriptions of the preferred embodiments of the present invention, and do not limit the scope of the present invention. Without departing from the design spirit of the present invention, various deformations and improvements made by those of ordinary skill in the art to the technical solutions of the present invention shall fall within the protection scope determined by the claims of the present invention.

Claims

1. A forest fire spread simulation method based on UAV LiDAR technology, characterized in that: The method comprises: Collect forest area data based on drone LiDAR technology, cameras and sensors; wherein the forest area data includes LiDAR point cloud data, thermal infrared data, image data and meteorological data; Based on the forest area data, extract forest area features to obtain feature parameters; Based on the characteristic parameters, a fire spread dynamic model is constructed through a fire spread dynamics reinforcement learning framework and a vertical-horizontal fire spread coupling algorithm; Based on the fire spread dynamic model, fire spread simulation and prediction are carried out, and dynamic early warning is carried out based on the simulation and prediction results to obtain fire fighting and fire prevention decisions.

2. The method according to claim 1, characterized in that The characteristic parameters include terrain parameters, fuel parameters and meteorological parameters; The terrain parameters include the altitude, slope and aspect of the forest area; The fuel parameters include surface fuel type, fuel moisture characteristics, surface fuel loading, canopy density, vertical fuel distribution, and canopy gap distribution; The meteorological parameters include temperature, humidity, wind speed, wind direction, atmospheric pressure and light intensity.

3. The method according to claim 2, characterized in that Methods for constructing a dynamic model of fire spread include: Based on terrain parameters, fuel parameters, and meteorological parameters, a deep reinforcement learning model is constructed, and the spread of surface fire and canopy fire is used as a Markov decision process to obtain the dynamic fire line position and intensity. Based on the flame heat radiation transfer equation and air turbulence model, combined with the canopy gap distribution, the flame jump distance and probability are calculated to obtain a coupling model of vertical spread and horizontal diffusion of canopy flames; Based on the coupling model, a three-dimensional fire propagation path is simulated and obtained; The fire spread dynamic model is constructed based on the dynamic fire line position and intensity and the three-dimensional fire propagation path.

4. The method according to claim 3, characterized in that: Methods for obtaining dynamic firing line position and intensity include: discretizing the terrain parameter, the fuel parameter, and the meteorological parameter; The spatial distribution characteristics of discretized terrain parameters and discretized fuel parameters are obtained based on convolutional neural networks, and the temporal characteristics of discretized meteorological parameters are captured based on long short-term memory networks. Based on the spatial distribution characteristics and the temporal characteristics, a state space is constructed; the position and intensity of the line of fire are used as an action space; Using the state space as the input of the deep Q network and the action space as the output of the deep Q network, updating the network weights by minimizing the loss function, and training the agent; The intelligent agent selects the optimal action according to the current state, simulates the Markov decision process, and obtains the dynamic firing line position and intensity.

5. The method according to claim 3, characterized in that: Methods for obtaining a coupled model of vertical and horizontal spread of canopy fire include: The flame heat radiation transmission equation is established by adopting the radiation heat transfer theory, and the propagation and attenuation process of flame heat radiation in the forest canopy is obtained based on the flame heat radiation transmission equation; Based on the air turbulence intensity, the air turbulence model is used to simulate the influence of air turbulence on flame spread; Based on the propagation and attenuation process of flame thermal radiation in the forest canopy and the influence of air turbulence on flame spread, combined with the canopy gap distribution, the flame jump distance and probability are calculated to obtain a coupling model of vertical spread and horizontal diffusion of forest canopy flames.

6. The method according to claim 3, characterized in that It also includes integrating LiDAR point cloud data, fire spread simulation and prediction results, and real-time meteorological data to establish a three-dimensional dynamic visualization platform, dynamically display the movement trend of the fire line through thermal maps and particle special effects, and divide the fire risk level.

7. A forest fire spread simulation system based on UAV LiDAR technology, used to implement the method described in any one of claims 1 to 6, characterized in that: include: A data acquisition module, used to collect forest area data based on UAV LiDAR technology, cameras and sensors; wherein the forest area data includes LiDAR point cloud data, thermal infrared data, image data and meteorological data; A feature parameter extraction module, used to extract forest area features based on the forest area data to obtain feature parameters; A fire spread dynamic model building module is used to build a fire spread dynamic model based on the characteristic parameters through a fire spread dynamics reinforcement learning framework and a vertical-horizontal fire spread coupling algorithm; The fire simulation module is used to simulate and predict the spread of fire based on the fire spread dynamic model, and to make dynamic warnings based on the simulation and prediction results to obtain fire fighting and fire prevention decisions.

8. The system according to claim 7, characterized in that In the characteristic parameter extraction module, the characteristic parameters include terrain parameters, fuel parameters and meteorological parameters; The terrain parameters include the altitude, slope and aspect of the forest area; The fuel parameters include surface fuel type, fuel moisture characteristics, surface fuel loading, canopy density, vertical fuel distribution, and canopy gap distribution; The meteorological parameters include temperature, humidity, wind speed, wind direction, atmospheric pressure and light intensity.

Images