A forest fire danger early warning system and method based on meteorological and vegetation data analysis

Abstract

The application discloses a forest fire risk early warning system and method based on meteorological and vegetation data analysis, relates to the technical field of fire risk prediction and control, and achieves the following effects: through high-resolution DEM and multi-spectral or hyperspectral images, a target forest region is finely divided into grids, and the vegetation type of each monitoring subunit is identified; combined with hourly wind speed, wind direction, relative humidity, precipitation and vapor pressure difference meteorological data, a wet-dry cycle index WDI, a monsoon effect index MFI and a dry thunderstorm risk factor DRF are calculated; through extraction of crown height, canopy density, undergrowth shrub and grass layer thickness and high VOC tree coverage, a fuel continuity coefficient FCI and a volatile flammable index VEI are calculated, accurate quantification of vertical ladder fuel and horizontal crown layer continuity and flammability is realized; a comprehensive fire risk index FFI is constructed and determined for hierarchical management, so that the dynamicity, hierarchy and pertinence of fire risk early warning are realized, and the forest fire risk prevention and control precision and the patrol resource utilization efficiency are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fire risk prediction and control, in particular to a forest fire risk early warning system and method based on meteorological and vegetation data analysis. BACKGROUND

[0002] In the existing forest fire risk management practice, most methods mainly rely on meteorological factors or a single vegetation index for fire risk assessment, and the focus is on emergency response and fire fighting after the occurrence of fire. For the stage when fire has not yet occurred, i.e. the prevention and control stage centered on patrol and preventive management, the existing technology still shows insufficient capability. Especially in monsoon climate regions, during the invasion or retreat of monsoon, wind speed and direction may change abruptly, leading to rapid drying of local dry fuel and increase of potential fire risk, but the traditional method fails to quantify the impact of such meteorological disturbance on fire risk potential and guide the scientific allocation of patrol resources.

[0003] In addition, the vertical fuel structure of the forest has an important influence on the potential spread path of fire. The thickness of the undergrowth layer, the distribution of crown height and the canopy density jointly determine the possibility of ground fire spreading upward to the canopy (ladder fuel effect) and the horizontal fire spread ability between the canopy. The existing method has shortcomings in quantifying the continuity of vertical fuel and the connectivity of horizontal fuel, and it is difficult to accurately identify high-risk forest stand units that have not yet occurred fire for key patrol and preventive management.

[0004] At the same time, the traditional fire risk monitoring system has low spatial resolution and lacks a priority patrol mechanism for high fire potential areas, resulting in low efficiency of patrol resource allocation and unsatisfactory preventive management effect. Therefore, there is an urgent need for a system that can identify high-risk areas through comprehensive analysis of meteorological and vegetation data, prioritize patrol and develop management strategies before the occurrence of fire, in order to improve the foresight and refinement level of forest fire risk prevention and control. SUMMARY

[0005] In view of the shortcomings of the prior art, the present application provides a forest fire risk early warning system and method based on meteorological and vegetation data analysis to solve the problems mentioned in the background.

[0006] To achieve the above purpose, the present application realizes the following technical scheme: a forest fire risk early warning system based on meteorological and vegetation data analysis, comprising:

[0007] A monitoring area division module is used to divide the target forest area into a plurality of monitoring sub-units according to a set spatial resolution based on a high-resolution digital elevation model DEM and remote sensing images, identify and extract the vegetation types in each monitoring sub-unit and mark them.

[0008] The meteorological monsoon characteristic analysis module is used to determine the monitoring date n and the latitude and longitude of the monitoring area based on GPS and astronomical epoch calculations. It divides the entire day into K time steps Δt and collects the hourly wind speed at the index of the kth time step. ,wind direction relative humidity Hourly precipitation With vapor pressure difference Establish a meteorological dataset; to construct:

[0009] The wet-dry cycle index (WDI) is used to characterize the rapid re-drying effect of fine fuels after heavy precipitation.

[0010] The Monsoon Effect Index (MFI) is used to characterize the exacerbating effect of sudden wind field changes during monsoon intrusion or retreat.

[0011] Furthermore, dry thunderstorm characteristics were identified, and a dry thunderstorm risk factor DRF was constructed, based on hourly lightning frequency. >5 times / hour, and hourly precipitation When the speed is less than 1 mm / h, it is determined to be a high ignition probability zone, and the first warning command is triggered;

[0012] The vegetation fuel characteristic analysis module is used to extract and calculate vegetation data within each monitoring sub-unit, including:

[0013] Canopy height distribution, canopy closure, and understory shrub and grass layer thickness are used to calculate the fuel continuity coefficient (FCI).

[0014] The proportion of combustible volatiles was extracted, including the coverage of pine, eucalyptus, rhododendron, and Lauraceae tree species, and a volatile flammability index (VEI) was constructed.

[0015] The fire risk comprehensive assessment module is used to construct the comprehensive fire risk index (FFI) for each monitoring sub-unit based on the wet-dry cycle index (WDI), monsoon effect index (MFI), dry thunderstorm risk factor (DRF), fuel continuity coefficient (FCI), and volatile flammability index (VEI). It then classifies the comprehensive fire risk index (FFI) of each monitoring sub-unit into risk levels, obtains the corresponding risk level and corresponding control strategies, and generates a first priority control group based on the first early warning instruction to prioritize the implementation of the corresponding control strategies.

[0016] Preferably, the monitoring area division module includes a data acquisition and preprocessing unit, a division unit, and a vegetation type identification unit;

[0017] The data acquisition and preprocessing unit is configured to set an external data interface for receiving or downloading multispectral or hyperspectral remote sensing image data from a satellite / airborne / unmanned aerial vehicle platform, and laser radar (LiDAR) point cloud and echo data from an airborne or ground platform, and performing radiation calibration, atmospheric correction, geometric correction, denoising and cloud and snow masking processing; performing ground / non-ground classification, outlier point removal and intensity normalization on the LiDAR point cloud; performing coordinate reference system and projection unification on the multispectral or hyperspectral remote sensing image data and the laser radar (LiDAR) point cloud and echo data, and performing spatio-temporal registration and co-location correction with a high-resolution digital elevation model (DEM), and resampling and cropping according to an initial spatial resolution R to generate a "vegetation data set" for subsequent analysis;

[0018] The division unit is configured to divide the target forest region into a set of monitoring sub-units in the DEM coordinate system at the initial spatial resolution R, and complete spatial labeling and indexing on the DEM.

[0019] The set of monitoring sub-units , to represent the 1st monitoring sub-unit to the Zth monitoring sub-unit.

[0020] The vegetation type identification unit is configured to perform classification by support vector machine after training based on the "vegetation data set" of the multispectral or hyperspectral image output by the data acquisition and preprocessing unit, and output the vegetation category of each monitoring sub-unit, the vegetation category including coniferous forest, broad-leaved forest, mixed forest, shrub and grassland.

[0021] Preferably, the vegetation type identification unit comprises a feature extraction unit and a classification unit.

[0022] The feature extraction unit is configured to extract spectral indicators based on the "vegetation data set" of the multispectral or hyperspectral image output by the data acquisition and preprocessing unit, the spectral indicators including normalized difference vegetation index (NDVI), red edge vegetation index (REVI) and short-wave infrared moisture index (MSI), and the expressions are as follows:

[0023]

[0024]

[0025]

[0026] wherein NIR is near-infrared band reflectivity, RED is red band reflectivity, RE is red edge band reflectivity, and SWIR is short-wave infrared band reflectivity.

[0027] A gray level co-occurrence matrix P(i,j) is constructed, and the specific steps are as follows: a spatial texture feature is extracted, specifically: the target band or principal component image is quantized to L=32 gray levels, the joint occurrence frequency of the gray level value of the neighborhood pixel of j is obtained by using a window size w=7x7 pixels, a distance d=1, and a direction theta epsilon {0, 45, 90, 135}, and the statistical result is normalized to obtain the gray level co-occurrence matrix P(i,j):

[0028]

[0029] wherein N(i,j) is the number of times of co-occurrence of the gray level i and the gray level j; i represents the gray level value of the reference pixel, j represents the gray level value of the neighborhood pixel of the reference pixel in the specified direction and distance, and P(i,j) represents the joint occurrence probability of the pixel with the gray level i and the neighborhood pixel with the gray level j in the window;

[0030] The texture indexes are calculated and averaged in the direction to obtain the texture indexes, including:

[0031] Contrast CON: ;

[0032] Homogeneity HOM: ;

[0033] Energy / angle second moment ASM: ;

[0034] Entropy ENT: ; wherein the index entropy is ignored when P(i,j)=0;

[0035] Correlation COR: wherein represents the row direction gray level mean value, represents the column direction gray level mean value, represents the row direction standard deviation, represents the column direction standard deviation,

[0036] A classification unit is used for classification based on the spectral indexes and the texture indexes, including:

[0037] When the NDVI value of the monitoring subunit is between 0.65 and 0.85, the REVI value is between 0.10 and 0.20, the MSI value is between 0.45 and 0.55, the texture index contrast CON is between 10 and 20, the homogeneity HOM is between 0.70 and 0.85, the energy ASM is between 0.55 and 0.70, the entropy ENT is between 3 and 4, and the correlation COR is between 0.85 and 0.95, it is determined that the monitoring subunit is a coniferous forest;

[0038] When the NDVI value of the monitoring subunit is between 0.60 and 0.80, the REVI value is between 0.15 and 0.25, the MSI value is between 0.50 and 0.60, the texture index CON is between 15 and 25, the HOM is between 0.60 and 0.75, the ASM is between 0.50 and 0.65, the ENT is between 3.5 and 4.5, and the COR is between 0.80 and 0.90, it is determined as broad-leaved forest;

[0039] When the NDVI value of the monitoring subunit is between 0.55 and 0.75, the REVI value is between 0.12 and 0.22, the MSI value is between 0.50 and 0.65, the texture index CON is between 18 and 28, the HOM is between 0.55 and 0.70, the ASM is between 0.45 and 0.60, the ENT is between 4 and 5, and the COR is between 0.75 and 0.85, it is determined as mixed forest;

[0040] When the NDVI value of the monitoring subunit is between 0.35 and 0.55, the REVI value is between 0.08 and 0.15, the MSI value is between 0.60 and 0.75, the texture index CON is between 22 and 35, the HOM is between 0.50 and 0.65, the ASM is between 0.40 and 0.55, the ENT is between 4.5 and 5.5, and the COR is between 0.70 and 0.80, it is determined as shrub;

[0041] When the NDVI value of the monitoring subunit is between 0.20 and 0.40, the REVI value is between 0.05 and 0.10, the MSI value is between 0.65 and 0.80, the texture index CON is between 12 and 20, the HOM is between 0.45 and 0.60, the ASM is between 0.35 and 0.50, the ENT is between 4 and 5, and the COR is between 0.65 and 0.75, it is determined as grassland.

[0042] Preferably, the meteorological monsoon feature analysis module comprises a meteorological data acquisition unit, a wet-dry cycle index calculation unit, and a monsoon effect index calculation unit;

[0043] The meteorological data acquisition unit is configured to acquire, in a monitoring date n, time step Δt, time moment index k = 1, …, K, the hourly wind speed , wind direction , relative humidity , hourly precipitation , and vapor pressure difference to establish a meteorological data set;

[0044] and decompose the scalar wind speed and wind direction into a two-dimensional vector: , ; the wind speed component in the east-west direction, the wind speed component in the north-south direction;

[0045] a wet-dry cycle index calculation unit for representing the effect of "strong rainfall rapidly moistens fine fuels, but the subsequent sunny and high vapor pressure difference leads to rapid re-drying within 1-2 days", which is calculated using a recent W-hour window, W=48h is set, and the wetness is calculated and the drying pull coefficient DI;

[0046]

[0047]

[0048] wherein W represents a recent W-hour set, is a scaling parameter, which is set to 10mm, ∈[0,1] represents the proximal wetness, the greater the rain, the closer to 1; is the number of effective time points in the window, when the research object is evapotranspiration, is the effective time point; it indicates that the atmosphere has certain water deficit conditions, and the plant leaf surface begins to be affected by the evapotranspiration pull, determines the "effective evapotranspiration period", that is, which time points in the statistical window are counted as effective samples; and , it indicates that the air is extremely dry, which exceeds the control ability of vegetation, and there is a risk of causing vegetation stomatal closure and even physiological stress; represents the reference vapor pressure difference risk identification threshold, which is set to 2kPa, and is used to determine "whether to enter the evapotranspiration stress state";

[0049] combined with the wetness and the drying pull coefficient DI, after non-dimensional treatment, the wet-dry cycle index WDI is calculated:

[0050] .

[0051] Preferably, a monsoon effect index calculation unit is used to represent the effect of wind field mutation and low-level jet intensification during the monsoon invasion / withdrawal period, when the wind speed / wind direction changes sharply, which is beneficial to fire spread and slope up, the wind vector difference is calculated to obtain the monsoon effect index MFI:

[0052]

[0053] wherein, the wind speed component in the east-west direction, the wind speed component in the north-south direction; and represent the change amount of the wind vector component of the adjacent two times.

[0054] Preferably, the meteorological monsoon feature analysis module further comprises a dry thunderstorm risk identification unit, the dry thunderstorm risk identification unit being configured to identify the hourly lightning frequency and the hourly precipitation amount , and preset a lightning threshold and a dry thunderstorm precipitation threshold , set to 5 times / hour; set to 1 mm / h;

[0055] define a thunderstorm risk factor DRF:

[0056]

[0057] In the monitoring date n, if the thunderstorm risk factor DRF = 1 at any time, the monitoring subunit corresponding to the monitoring date is marked as a 'high ignition probability area', and a first warning instruction is triggered, indicating that the monitoring subunit has a risk of lightning igniting dry vegetation or surface combustible materials, and through grid aggregation, a dry thunderstorm risk layer is output.

[0058] Preferably, the vegetation fuel feature analysis module comprises a canopy height distribution extraction unit, a canopy density calculation unit, and an understory shrub and grass layer thickness determination unit.

[0059] The canopy height distribution extraction unit is configured to extract a canopy height distribution curve within the monitoring subunit using LiDAR point cloud data or a forest three-dimensional point cloud model constructed from stereo image pairs, and calculate the average canopy height Hmean, the maximum canopy height Hmax, and the standard deviation σh of the height distribution:

[0060]

[0061] wherein N is the total number of point clouds in the monitoring subunit, is the height of the pth point;

[0062] The canopy density calculation unit is configured to calculate the canopy density C of the monitoring subunit based on the remote sensing image in the "vegetation data set" through point cloud gap rate analysis technology, the range of which is 0~1, representing the proportion of the ground covered by the canopy, and the calculation formula of the canopy density C is:

[0063]

[0064] wherein GapArea is the visible gap area of the ground, and TotalArea is the total area of the monitoring subunit;

[0065] The understory shrub and grass layer thickness determination unit is configured to calculate the average thickness Hunder of the shrub and herbaceous vegetation within the monitoring subunit based on the low-layer point cloud echo in the "vegetation data set" of the hyperspectral image, and the calculation formula is:

[0066]

[0067] wherein M is the number of low-level pixels in the monitoring subunit, is the height of the qth low-level pixel, and the low-level includes the height between 0-2m from the ground.

[0068] Preferably, the vegetation fuel feature analysis module further comprises a fuel continuity analysis unit and a combustible species identification unit;

[0069] The fuel continuity analysis unit is used to normalize and boundary process the average canopy height Hmean, the maximum canopy height Hmax, and the standard deviation of height distribution σh, the canopy density C, and the average thickness of shrubs and herbaceous vegetation Hunder, to obtain the relative uniformity of canopy m, the vertical stratification intensity of canopy s, the relative amount of "ladder fuel" thickness under the canopy u, and the horizontal canopy density c, by using the clipping function.

[0070]

[0071]

[0072]

[0073]

[0074] The horizontal continuity component is constructed to obtain the horizontal continuity index Hhor:

[0075]

[0076] The vertical continuity component is constructed to obtain the vertical continuity index Hver:

[0077]

[0078] wherein, and are weights, , , the thickness under the canopy is slightly heavier than the stratification degree;

[0079] and the fuel continuity coefficient FCI is calculated by combining the horizontal continuity index Hhor and the vertical continuity index Hver:

[0080]

[0081] wherein, and are weights, , ;

[0082] A combustible species identification unit is used to extract high VOC tree species coverage in each monitoring subunit, including: pine, eucalyptus, rhododendron, and laurel tree species coverage, and to construct a volatile flammable index VEI for each monitoring subunit:

[0083]

[0084] wherein, represents the number of high VOC tree species, represents the volatile weight coefficient of the a-th tree species, wherein pine is 0.8, eucalyptus is 0.95, rhododendron is 0.7, and laurel is 0.6; represents the coverage of the a-th tree species in the r-th monitoring subunit.

[0085] Preferably, the fire risk comprehensive assessment module includes a comprehensive fitting unit and a judgment unit;

[0086] The comprehensive fitting unit is used to extract the wet-dry cycle index WD I, the monsoon effect index MFI, the dry thunderstorm risk factor DRF, the fuel continuity coefficient FCI, and the volatile flammable index VEI of each monitoring subunit, and after normalization, a comprehensive fire risk index FFI is constructed for each monitoring subunit:

[0087]

[0088]

[0089] wherein, , , , , is the weight;

[0090] The judgment unit is used to classify the risk level of the comprehensive fire risk index FFI of each monitoring subunit, specifically:

[0091] When FFI>0.7, the label is the first risk level, and a first control strategy is generated, including: checking once a day, prohibiting all open fires, including barbecues, burning waste, and fireworks and firecrackers; and setting obvious warning signs along the forest area and management road, at least one "strictly prohibit open fire" sign every 500 meters; cleaning the 10-meter-wide strip area along the patrol route, forest roads, and forest edges within a 50-meter range of dry branches, leaves, hay stacks, and waste wood at least once a week;

[0092] When 0.4≤FFI≤0.7, the label is the second risk level, and a second control strategy is generated, including: checking once every two days, and setting at least one "strictly prohibit open fire" sign every 800 meters; cleaning the 10-meter-wide strip area along the patrol route, forest roads, and forest edges within a 30-meter range of dry branches, leaves, hay stacks, and waste wood once every two weeks;

[0093] When FFI < 0.4, the label is the third risk level, and the third control strategy is generated, including: weekly patrol once, setting 1 "strictly prohibited open fire" sign every 1000 meters, cleaning 3-5 meters wide strip area along the patrol route and 10 meters range of forest road, forest edge, dry branches, leaves, hay stacks and waste wood every month;

[0094] When the monitoring sub-unit is determined as a 'high ignition probability area' and the first early warning instruction is triggered, the monitoring sub-unit of the first early warning instruction is counted to form a first priority control group, and the corresponding control strategy is preferentially implemented.

[0095] Preferably, the following steps are included:

[0096] S1, obtaining a high-resolution digital elevation model DEM and multispectral or hyperspectral remote sensing image, dividing the target forest area into a plurality of monitoring sub-units according to a set spatial resolution, and identifying and marking the vegetation type in each monitoring sub-unit;

[0097] S2, determining the monitoring date n and the latitude and longitude of the monitoring area based on global positioning system GPS and astronomical epoch calculation, dividing the whole day into K time steps Δt, collecting the hourly wind speed , wind direction , relative humidity , hourly precipitation and vapor pressure difference at the kth time index, and establishing a weather data set; constructing a wet-dry cycle index WD I, a monsoon effect index MFI and a dry thunderstorm risk factor DRF, if the thunderstorm risk factor DRF = 1 at any time, the monitoring sub-unit corresponding to the monitoring date is marked as a 'high ignition probability area', and the first early warning instruction is triggered, indicating that the monitoring sub-unit has the risk of lightning igniting dry vegetation or surface combustible material, and through grid aggregation, a dry thunderstorm risk layer is output;

[0098] S3, extracting vegetation data in each monitoring sub-unit, and calculating fuel continuity coefficient FCI and volatile flammable index VEI;

[0099] S4, based on the wet-dry cycle index WD I, the monsoon effect index MFI, the dry thunderstorm risk factor DRF, the fuel continuity coefficient FCI and the volatile flammable index VEI, the comprehensive fire risk index FFI of each monitoring sub-unit is constructed, and the comprehensive fire risk index FFI of each monitoring sub-unit is classified by risk level to obtain the corresponding risk level and the corresponding control strategy, and combined with the first early warning instruction, a first priority control group is generated to preferentially implement the corresponding control strategy.

[0100] The application provides a forest fire risk early warning system and method based on meteorological and vegetation data analysis.

[0101] (1) The application realizes fine quantitative evaluation of fire risk potential by comprehensively considering monsoon climate conditions and forest vertical fuel structure, effectively solving the problem that the traditional fire risk monitoring method only relies on meteorological factors or a single vegetation index and is difficult to reflect the risk of rapid spread of fire. The application establishes a wet-dry cycle index WD I and a monsoon effect index MFI, which can quantify the re-drying speed of fine fuel after heavy rain and the intensifying effect of wind speed and direction mutation during monsoon invasion or retreat on fire spread, thereby accurately identifying the potential risk of local rapid spread of fire. At the same time, by extracting and analyzing the crown height distribution, canopy density and understory shrub layer thickness of each monitoring subunit, calculating the fuel continuity coefficient FCI, and combining the flammable volatile matter ratio to construct the volatile matter flammability index VEI, the vertical fuel gradient and horizontal connectivity influence are quantitatively characterized, which can accurately evaluate the potential path of ground fire spreading upward to the forest canopy and lateral diffusion between the crown layers, fully reflecting the ladder fuel effect and crown connectivity effect. Further, the application constructs a dry thunderstorm risk factor DRF and combines the first early warning instruction to include the high ignition probability area into the first priority control group, realizing the priority patrol and control of high-risk areas, significantly improving the utilization efficiency of patrol resources and the effect of fire risk prevention and control. At the same time, the application uses high-resolution DEM and remote sensing images to divide the monitoring area into regular grids, accurately extracts the vegetation type, realizes the zoning and grading management of forest stand heterogeneity in space, makes the fire risk warning be able to implement differentiated control strategies for different subunits, effectively reduces the suddenness and uncontrollability of fire, and enhances the scientificity, accuracy and operability of forest fire risk prevention and control.

[0102] (2) The wet-dry cycle index and the monsoon effect index are calculated using hourly meteorological data, which can dynamically reflect the influence of precipitation re-drying and wind field mutation on fire risk potential, realize meteorological-driven dynamic fire risk warning, and improve the timeliness of prevention and control.

[0103] (3) The fuel continuity coefficient and the volatile matter flammability index quantify the crown connectivity, vertical ladder fuel thickness and high VOC species distribution, which can scientifically evaluate the horizontal spread and vertical ignition potential of fire, and assist in formulating targeted control strategies.

[0104] (4) The comprehensive fire risk index FFI combined with the priority identification of high ignition probability area realizes the hierarchical management and optimal allocation of key patrol resources, effectively reduces the fire occurrence rate and rapid spread risk, and improves the efficiency of forest fire prevention and control. BRIEF DESCRIPTION OF DRAWINGS

[0105] Figure 1 is a schematic diagram of the system flow of the application;

[0106] Figure 2 This is a flowchart illustrating the monitoring area division module of the present invention;

[0107] Figure 3 This is a flowchart illustrating the meteorological monsoon characteristic analysis module of the present invention;

[0108] Figure 4 This is a flowchart illustrating the vegetation fuel characteristic analysis module of the present invention.

[0109] Figure 5 This is a schematic diagram of the fire risk comprehensive assessment module of the present invention;

[0110] Figure 6 This is a schematic diagram of the method steps of the present invention. Detailed Implementation

[0111] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0112] Example 1

[0113] Please see Figures 1 to 5 This invention provides a forest fire risk early warning system based on meteorological and vegetation data analysis, comprising:

[0114] The monitoring area division module is used to divide the target forest area into several monitoring sub-units according to a set spatial resolution by regular grid division based on high-resolution digital elevation model (DEM) and remote sensing imagery, and to identify and extract the vegetation type in each monitoring sub-unit and mark it.

[0115] The meteorological monsoon characteristic analysis module is used to determine the monitoring date n and the latitude and longitude of the monitoring area based on GPS and astronomical epoch calculations. It divides the entire day into K time steps Δt and collects the hourly wind speed at the index of the kth time step. ,wind direction relative humidity Hourly precipitation With vapor pressure difference Establish a meteorological dataset; to construct:

[0116] The wet-dry cycle index (WDI) is used to characterize the rapid re-drying effect of fine fuels after heavy precipitation.

[0117] The Monsoon Effect Index (MFI) is used to characterize the exacerbating effect of sudden wind field changes during monsoon intrusion or retreat.

[0118] Further identify dry thunderstorm characteristics, build dry thunderstorm risk factor DRF, when the hourly lightning frequency >5 times / hour, and the hourly precipitation <1mm / h, determine as high ignition probability area, and trigger the first early warning instruction;

[0119] Vegetation fuel characteristic analysis module, for extracting and calculating vegetation data in each monitoring subunit, including:

[0120] Canopy height distribution, canopy density, understory shrub and grass layer thickness, for calculating fuel continuity coefficient FCI;

[0121] Extract the proportion of flammable volatile matter, including the coverage rate of pine, eucalyptus, azalea and laurel, and build volatile matter flammability index VEI;

[0122] Fire risk comprehensive assessment module, for building comprehensive fire risk index FFI of each monitoring subunit based on wet-dry cycle index WD I, monsoon effect index MFI, dry thunderstorm risk factor DRF, fuel continuity coefficient FCI and volatile matter flammability index VEI, and grading the comprehensive fire risk index FFI of each monitoring subunit to obtain the corresponding risk level and the corresponding control strategy, and generate the first priority control group to implement the corresponding control strategy preferentially in combination with the first early warning instruction.

[0123] In this embodiment, the present application realizes fine quantitative evaluation of fire danger potential by comprehensively considering monsoon climate conditions and forest vertical fuel structure, effectively solving the problem that traditional fire danger monitoring methods only rely on meteorological factors or a single vegetation index, and are difficult to reflect the risk of rapid fire spread. The present application establishes a wet-dry cycle index WDI and a monsoon effect index MFI, which can quantify the re-drying speed of fine fuel after heavy rain and the intensifying effect of wind speed and direction mutation during monsoon invasion or retreat on fire spread, thereby accurately identifying the potential risk of local rapid fire spread. At the same time, by extracting and analyzing the crown height distribution, canopy density and understory shrub layer thickness of each monitoring subunit, calculating the fuel continuity coefficient FCI, and combining the flammable volatile matter ratio to construct the volatile matter flammability index VEI, the vertical fuel gradient and horizontal connectivity influence are quantitatively characterized, which can accurately evaluate the potential path of surface fire spreading to the forest canopy and lateral diffusion of crown layer fire, fully reflecting the ladder fuel effect and crown layer connectivity effect. Further, the present application constructs a dry thunderstorm risk factor DRF and combines the first warning instruction to include the high fire probability area into the first priority control group, realizing the priority patrol and control of high-risk areas, significantly improving the utilization efficiency of patrol resources and the effect of fire danger prevention and control. At the same time, the present application uses high-resolution DEM and remote sensing images to divide the monitoring area into regular grids, accurately extracts the vegetation type, realizes the zoning and grading management of forest stand heterogeneity in space, so that the fire danger warning can implement differentiated control strategies for different subunits, effectively reducing the suddenness and uncontrollability of fire, and enhancing the scientificity, accuracy and operability of forest fire danger prevention and control.

[0124] Embodiment 2

[0125] This embodiment is an explanation and description in embodiment 1, please refer to Figures 1 to 5 , specifically, the monitoring area division module includes a data acquisition and preprocessing unit, a division unit, and a vegetation type identification unit;

[0126] The data acquisition and preprocessing unit is used to set an external data interface for receiving or downloading multispectral or hyperspectral remote sensing image data from satellite / airborne / unmanned aerial vehicle platforms, and laser radar LiDAR point cloud and echo data from airborne or ground platforms, and performing radiation calibration, atmospheric correction, geometric correction, denoising and cloud and snow masking processing; performing ground / non-ground classification, outlier rejection and intensity normalization on the LiDAR point cloud; performing coordinate reference system and projection unification on the multispectral or hyperspectral remote sensing image data and the laser radar LiDAR point cloud and echo data, and performing time and space registration and same-site correction with the high-resolution digital elevation model DEM, and resampling and cropping according to the initial spatial resolution R, to generate a "vegetation data set" for subsequent analysis;

[0127] A dividing unit is configured to divide the target forest region into a set of monitoring sub-units in a DEM coordinate system at an initial spatial resolution R, preferably 10 m x 10 m, and complete spatial labeling and indexing on the DEM;

[0128] the set of monitoring sub-units , to represent the 1st monitoring sub-unit to the Zth monitoring sub-unit;

[0129] A vegetation type identifying unit is configured to train a support vector machine based on a “vegetation data set” of multi-spectral or hyperspectral images output by the data acquisition and preprocessing unit, and then perform classification by the support vector machine to output a vegetation category of each monitoring sub-unit, the vegetation category including coniferous forest, broad-leaved forest, mixed forest, shrub, and grassland.

[0130] In this embodiment, the monitoring area dividing module is refined into a data acquisition and preprocessing unit, a dividing unit, and a vegetation type identifying unit, achieving high-precision and systematic management of the forest region, thereby significantly improving the accuracy and scientificity of fire risk early warning. In the data acquisition and preprocessing unit, multi-spectral / hyperspectral remote sensing images of satellite, airborne, and unmanned aerial vehicle platforms, as well as LiDAR point cloud and echo data, are integrated, and radiation calibration, atmospheric correction, geometric correction, denoising, cloud and snow masking processing, ground / non-ground classification, outlier rejection, and intensity normalization are performed to ensure the high quality and comparability of the original data. At the same time, through coordinate system unification, projection registration, and same-site correction with high-resolution DEM, accurate spatio-temporal matching of multi-source data is achieved, providing a reliable basis for subsequent analysis. The dividing unit divides the target forest region into a set of monitoring sub-units according to the preferred spatial resolution (e.g., 10 m x 10 m), and completes spatial labeling and indexing on the DEM, enabling fine expression of forest heterogeneity. The vegetation type identifying unit trains and classifies the “vegetation data set” using a supervised classification method and a support vector machine, achieving accurate identification of different vegetation types such as coniferous forest, broad-leaved forest, mixed forest, shrub, and grassland, and enabling each monitoring sub-unit to obtain clear vegetation category information. This method can fully reflect the influence of different stand structures and vertical fuel characteristics on fire risk, providing a scientific basis for comprehensive assessment and differentiated control of fire risk, significantly improving the precision and efficiency of forest fire risk early warning, optimizing patrol resource allocation, and reducing the risk of fire outbreak and spread.

[0131] Embodiment 3

[0132] This embodiment is an explanation and description in Embodiment 2, please refer to Figures 1 to 5 , specifically, the vegetation type identifying unit includes a feature extraction unit and a classification unit;

[0133] The feature extraction unit is configured to extract spectral indexes based on the "vegetation data set" of the multispectral or hyperspectral image output by the data acquisition and preprocessing unit, the spectral indexes including a normalized difference vegetation index NDVI, a red edge vegetation index REVI, and a short-wave infrared moisture index MSI, and the expressions are as follows:

[0134]

[0135]

[0136]

[0137] wherein NIR is the reflectivity of the near-infrared band, RED is the reflectivity of the red band, RE is the reflectivity of the red edge band, and SWIR is the reflectivity of the short-wave infrared band;

[0138] A gray level co-occurrence matrix P(i,j) is constructed, and the specific steps are as follows: spatial texture features are extracted, specifically, the target band or principal component image is quantized to L=32 gray levels, the joint occurrence frequency of the gray level values of the neighborhood pixels of the window size w=7x7 pixels, distance d=1, and direction theta belonging to {0°, 45°, 90°, 135°} is j, the statistical results are normalized to obtain the gray level co-occurrence matrix P(i,j):

[0139]

[0140] wherein N(i,j) is the number of co-occurrences of the gray levels i and j; i represents the gray level value of the reference pixel, j represents the gray level value of the neighborhood pixel of the reference pixel in the specified direction and distance, and P(i,j) represents the joint occurrence probability of the pixel with the gray level i and the neighborhood pixel with the gray level j in the window;

[0141] The texture indexes are calculated and averaged in the direction to obtain the texture indexes, including:

[0142] Contrast CON: ; which represents the degree of difference in the gray level of the pixel, and the greater the value, the rougher the texture;

[0143] Homogeneity HOM: ; which represents the concentration of the diagonal adjacent elements of the pixel, and the greater the value, the more uniform the texture;

[0144] Energy / Angular Second Moment ASM: ; which represents the uniformity of the gray level distribution, and the greater the value, the more regular the texture of the image;

[0145] Entropy ENT: ; wherein the index entropy is ignored when P(i,j)=0. which represents the complexity of the texture, and the greater the value, the more irregular the texture;

[0146] Correlation COR: wherein, represents the mean of the row direction gray level, represents the mean of the column direction gray level, represents the standard deviation of the row direction, represents the standard deviation of the column direction, COR represents the degree of linear dependence between the pixels, and a value close to 1 indicates that the texture gray level is highly correlated;

[0147] a classification unit for classifying based on the spectral index and the texture index, comprising:

[0148] when the NDVI value of the monitoring subunit is between 0.65 and 0.85, the REVI value is between 0.10 and 0.20, the MSI value is between 0.45 and 0.55, and the texture index contrast CON is between 10 and 20, the homogeneity HOM is between 0.70 and 0.85, the energy ASM is between 0.55 and 0.70, the entropy ENT is between 3 and 4, and the correlation COR is between 0.85 and 0.95, it is determined that the monitoring subunit is a coniferous forest;

[0149] when the NDVI value of the monitoring subunit is between 0.60 and 0.80, the REVI value is between 0.15 and 0.25, the MSI value is between 0.50 and 0.60, and the texture index CON is between 15 and 25, the HOM is between 0.60 and 0.75, the ASM is between 0.50 and 0.65, the ENT is between 3.5 and 4.5, and the COR is between 0.80 and 0.90, it is determined to be a broad-leaved forest;

[0150] when the NDVI value of the monitoring subunit is between 0.55 and 0.75, the REVI value is between 0.12 and 0.22, the MSI value is between 0.50 and 0.65, and the texture index CON is between 18 and 28, the HOM is between 0.55 and 0.70, the ASM is between 0.45 and 0.60, the ENT is between 4 and 5, and the COR is between 0.75 and 0.85, it is determined to be a mixed forest;

[0151] when the NDVI value of the monitoring subunit is between 0.35 and 0.55, the REVI value is between 0.08 and 0.15, the MSI value is between 0.60 and 0.75, and the texture index CON is between 22 and 35, the HOM is between 0.50 and 0.65, the ASM is between 0.40 and 0.55, the ENT is between 4.5 and 5.5, and the COR is between 0.70 and 0.80, it is determined to be a shrub;

[0152] When the NDVI value of the monitoring subunit is between 0.20 and 0.40, the REVI value is between 0.05 and 0.10, the MSI value is between 0.65 and 0.80, the texture index CON is between 12 and 20, the HOM is between 0.45 and 0.60, the ASM is between 0.35 and 0.50, the ENT is between 4 and 5, and the COR is between 0.65 and 0.75, it is determined to be grassland. The vegetation type classification table is shown in Table 1 below.

[0153] Table 1: Vegetation type classification table

[0154]

[0155] In this embodiment, the feature extraction unit and the classification unit are introduced to realize high-precision and automatic identification of the vegetation type in the monitoring subunit, thereby significantly improving the scientificity and pertinence of forest fire danger early warning. In the feature extraction unit, the spectral indexes (such as NDVI, REVI, and MSI) and the gray level co-occurrence matrix texture features (including contrast CON, homogeneity HOM, energy ASM, entropy ENT, and correlation COR) of multispectral or hyperspectral images are comprehensively utilized, which not only represents the growth condition, leaf green coverage, and water content of vegetation, but also reflects the vertical structure and spatial heterogeneity of forest, providing multi-dimensional information for fine differentiation of different vegetation types such as coniferous forest, broad-leaved forest, mixed forest, shrub, and grassland. In the classification unit, by setting multi-dimensional thresholds of spectral indexes and texture indexes, accurate classification based on rules is realized, so that the vegetation type of the monitoring subunit can be clearly labeled. This method not only reflects the influence of vertical and horizontal differences of stand structure on fire danger, but also provides a reliable basis for fuel continuity assessment and fire danger index calculation, effectively improving the precision of comprehensive fire danger assessment and the pertinence of patrol strategy, and optimizing resource allocation.

[0156] Embodiment 4

[0157] This embodiment is an explanation and description in Embodiment 1, please refer to Figures 1 to 5 , specifically, the meteorological monsoon feature analysis module includes a meteorological data acquisition unit, a wet-dry cycle index calculation unit, and a monsoon effect index calculation unit;

[0158] The meteorological data acquisition unit is used to acquire hourly wind speed (m / s), wind direction (°), relative humidity (%), hourly precipitation (mm), and vapor pressure difference (kPa) within the monitoring date n with time step Δt to form time index k = 1, …, K, and establish a meteorological data set;

[0159] And the scalar wind speed and wind direction are decomposed into two-dimensional vectors: , ; represents the wind speed component in the east-west direction, represents the wind speed component in the north-south direction;

[0160] The wet-dry cycle index calculation unit is used to represent the effect of "strong rainfall rapidly moistens fine fuels, but the subsequent sunny and high vapor pressure difference leads to rapid re-drying within 1-2 days", and is calculated using a recent W-hour window, with W=48h, to calculate the wetness and the drying pull coefficient DI;

[0161]

[0162]

[0163] where W represents a recent W-hour set, is a scaling parameter, set to 10mm, ∈[0,1] represents the near-end wetness, the greater the rain, the closer to 1; is the number of effective times in the window, when the research object is evapotranspiration, is the effective time; it means that the atmosphere has certain water deficit conditions, and the plant leaf surface starts to be affected by the evapotranspiration pull, and determines the "effective evapotranspiration period", that is, which time in the statistical window is counted as an effective sample; and , it indicates that the air is extremely dry, and the atmospheric evapotranspiration pull is strong, often exceeding the vegetation control capacity, and there is a risk of causing vegetation stomatal closure and even physiological stress; represents the reference vapor pressure difference risk identification threshold, set to 2kPa, used to determine "whether to enter the evapotranspiration stress state"; DI reflects the strength of the near-end drying driving force;

[0164] Combined with the wetness and the drying pull coefficient DI, after dimensionless processing, the wet-dry cycle index WDI is calculated:

[0165] .

[0166] The meaning of the formula is: when "there is significant rainfall near the end" ( is higher) and "the VPD is high in the subsequent 48h" (DI is larger), it means that there is a "wet-dry cycle after which it is more flammable";

[0167] Example 5

[0168] This embodiment is an explanation and description in Example 4, please refer to Figures 1 to 5, specifically, the monsoon effect index calculation unit is used to represent the role of wind field mutation and low-level jet intensification during the monsoon invasion / withdrawal period, when the wind speed / direction changes sharply, which is conducive to fire spread and uphill movement. The wind vector difference is calculated to obtain the monsoon effect index MFI:

[0169]

[0170] wherein, represents the wind speed component in the east-west direction, represents the wind speed component in the north-south direction; and represents the change amount of wind vector components of adjacent two times; if MFI is larger, it means that the wind direction / speed changes sharply between adjacent moments (or levels) → the wind field is unstable or fluctuates greatly; if MFI is smaller, it means that the wind direction / speed between adjacent moments is relatively smooth → the wind field is relatively stable.

[0171] The meteorological monsoon feature analysis module obtains hourly wind speed, wind direction, relative humidity, precipitation and vapor pressure difference and other meteorological parameters, and decomposes the wind speed and wind direction into two-dimensional vectors, providing accurate dynamic meteorological basic data for subsequent fire danger analysis; the wet-dry cycle index calculation unit can quantitatively represent the effect of rapid fuel wetting by heavy precipitation and subsequent high vapor pressure difference and sunny weather leading to rapid fuel drying, by calculating the wetness and drying pull coefficient DI, identifying the effective evapotranspiration period and determining the evapotranspiration stress state, realizing the sensitive capture of the influence of short-term wet-dry cycle on fire danger; when recent precipitation is significant and atmospheric evapotranspiration pull is enhanced within the next 48 hours, WDI value increases, which can effectively reflect the potential risk of "fuel more flammable after wet-dry cycle", thereby providing quantitative judgment on fire danger potential. This method overcomes the limitations of traditional fire danger warning relying only on precipitation or temperature and humidity single factor, and can dynamically and finely evaluate local fire danger risk in monsoon climate region, providing scientific basis for fire danger classification, patrol scheduling and priority control.

[0172] The meteorological monsoon feature analysis module further comprises a dry thunderstorm risk identification unit, which is used to identify the hourly lightning frequency and the hourly precipitation , and preset the lightning threshold and the dry thunderstorm precipitation threshold , set to 5 times / hour; set to 1 mm / h;

[0173] Define the thunderstorm risk factor DRF:

[0174]

[0175] Physical meaning, when lightning is more but precipitation is weak, it means that the atmospheric convection is strong enough to trigger lightning, but the raindrops are evaporated in the high temperature and low humidity environment during the falling process (i.e. virga phenomenon), and finally cannot effectively increase the surface humidity → the ignition potential increases;

[0176] In the monitoring date n, if the thunderstorm risk factor DRF = 1 at any time, the monitoring subunit corresponding to the monitoring date is marked as 'high ignition probability area', and the first warning instruction is triggered, indicating that the monitoring subunit has the risk of lightning igniting dry vegetation or surface combustible material, and through grid aggregation, a dry thunderstorm risk layer is output.

[0177] In this embodiment, the dry thunderstorm risk identification unit realizes the identification of dry thunderstorm events by monitoring the hourly lightning frequency and hourly precipitation, and setting the lightning threshold to 5 times / hour and the dry thunderstorm precipitation threshold to 1 mm / h; define the thunderstorm risk factor DRF, when DRF = 1, it means that lightning is frequent but precipitation is weak, which physically reflects that the atmospheric convection is strong enough to trigger lightning, but the raindrops are evaporated in the high temperature and low humidity environment during the falling process (virga phenomenon), which cannot effectively increase the surface humidity, thereby significantly increasing the ignition potential of the surface combustible material and dry vegetation; In the monitoring date n, if DRF = 1 at any time, the corresponding monitoring subunit is marked as 'high ignition probability area', the first warning instruction is triggered, and a dry thunderstorm risk layer is generated through grid aggregation to realize the visualization and spatial distribution evaluation of the high ignition potential area. This method can timely and accurately identify the high-risk area of fire danger caused by dry thunderstorm, make up for the deficiency of traditional fire danger warning in identifying local strong convective weather, provide a scientific basis for the implementation of priority patrol and control strategy, and effectively improve the pertinence and efficiency of forest fire danger prevention and control.

[0178] Embodiment 6

[0179] This embodiment is an explanation and description in embodiment 1, please refer to Figures 1 to 5 , Specifically, the vegetation fuel characteristic analysis module includes a canopy height distribution extraction unit, a canopy density calculation unit and an understory shrub layer thickness determination unit;

[0180] The canopy height distribution extraction unit is used to extract the canopy height distribution curve in the monitoring subunit by using the LiDAR point cloud data or the three-dimensional point cloud model of the forest constructed by stereo image pairs, and calculate the average canopy height Hmean, the maximum canopy height Hmax and the standard deviation σh of the height distribution:

[0181]

[0182] Wherein, N is the total number of point clouds in the monitoring subunit, is the height of the pth point;

[0183] A canopy density calculation unit is configured to calculate a canopy density C of the monitoring subunit based on remote sensing images and by using a point cloud gap ratio analysis technique, wherein the canopy density C ranges from 0 to 1 and represents a proportion of a ground surface covered by a canopy, and a calculation formula of the canopy density C is:

[0184]

[0185] wherein GapArea is an area of a ground surface visible gap, and TotalArea is a total area of the monitoring subunit;

[0186] An undergrowth shrub and grass layer thickness determination unit is configured to calculate an average thickness Hunder of shrubs and herbaceous vegetation in the monitoring subunit based on low-layer point cloud echoes or high-resolution image texture features, and a calculation formula is:

[0187]

[0188] wherein M is a number of low-layer pixels in the monitoring subunit, is a height of a qth low-layer pixel, and the low layer includes a height between 0-2 m from the ground.

[0189] In this embodiment, fine extraction and analysis of vertical structures of vegetation in the monitoring subunit are performed to realize quantitative evaluation of fire risk potential. A canopy height distribution extraction unit constructs a three-dimensional forest point cloud model by using LiDAR point clouds or stereo images, extracts a canopy height distribution curve in the monitoring subunit, and calculates an average canopy height Hmean, a maximum canopy height Hmax, and a height distribution standard deviation σh, which can accurately represent a potential path of ground fire spreading upward to a canopy and a possibility of vertical fire spread between canopies. A canopy density calculation unit calculates a canopy density C based on remote sensing images and a point cloud gap ratio analysis technique, which reflects a proportion of a ground surface covered by a canopy and provides a basis for evaluating a spreading ability of fire in a horizontal plane. An undergrowth shrub and grass layer thickness determination unit calculates an average thickness Hunder of shrubs and herbaceous vegetation in a height range of 0-2 m by using low-layer point cloud echoes or high-resolution image texture features, which helps to quantify a "ladder fuel" effect, i.e., a risk of ground fire spreading upward to a canopy along low-layer vegetation. This module can comprehensively and finely represent vertical and horizontal fuel structures of a forest, provide scientific data support for comprehensive fire risk evaluation, make up for a lack of quantitative analysis of forest heterogeneity and vertical fuel continuity in traditional methods, and improve accuracy and pertinence of fire risk early warning.

[0190] Embodiment 7

[0191] This embodiment is an explanation and description made in Embodiment 6, please refer to Figures 1 to 5 Specifically, the vegetation fuel feature analysis module includes a fuel continuity analysis unit and a combustible species identification unit.

[0192] A fuel continuity analysis unit is configured to obtain a relative uniformity m of a canopy, a vertical stratification intensity s of the canopy, a relative amount u of an undergrowth "ladder fuel" thickness, and a horizontal canopy density c by using a clipping function to normalize and boundary process an average canopy height Hmean, a maximum canopy height Hmax, a standard deviation σh of a height distribution, a canopy density C, and an average thickness Hunder of shrubs and herbaceous vegetation.

[0193]

[0194]

[0195]

[0196]

[0197] wherein the clipping function clip(x, 0, 1) = min(max(x, 0), 1); m, s, and u can be uniformly set to 0 and processed separately for shrub / grassland scenarios to avoid abnormality caused by division of a minimum value;

[0198] Physical meaning, m: relative uniformity of the canopy (proportion of the average canopy height to the maximum canopy height). The closer to 1, the more uniform the canopy, and the more "uniform" the canopy top, which is conducive to crown-to-crown horizontal spread. s: vertical stratification intensity of the canopy (relative amount of height dispersion). The greater, the more layers and the more obvious the ladder, which is conducive to vertical connection of the ground-undergrowth-canopy ("ladder fuel").

[0199] u: relative amount of undergrowth "ladder fuel" thickness. The greater, the easier to "bridge" between the ground and the canopy by the combustible, which is conducive to vertical upward fire. c: horizontal connectivity / canopy density. The greater, the more continuous the canopy horizontally, which is conducive to horizontal spread of crown fire. If Hmax < 1 m (shrub / grass / bare land scenario), m = s = u = 0;

[0200] A horizontal continuity component is constructed to obtain a horizontal continuity index Hhor:

[0201]

[0202] wherein a high canopy coverage (large c) and a more uniform canopy height (large m) are more conducive to horizontal spread of crown-to-crown connection; any deficiency will weaken the probability of continuous spread of crown fire (multiplication suppresses the "short board effect");

[0203] A vertical continuity component is constructed to obtain a vertical continuity index Hver:

[0204]

[0205] Wherein, u: the thicker the understory layer, the easier it is to introduce the surface fire to the tree crown; s: the greater the crown height dispersion, the more layers, the easier it is to form a continuous vertical combustible band; and is a weight, , , the understory thickness is slightly heavier than the layering degree;

[0206] Combined with the horizontal continuity index Hhor and the vertical continuity index Hver, the fuel continuity coefficient FCI is calculated:

[0207]

[0208] wherein, and is a weight, , In most forest fire behaviors, horizontal crown layer continuity has a more direct impact on large-scale spread speed, so ; FCI∈[0, 1], the greater the value, the more continuous the space fuel, the higher the fire propagation potential.

[0209] Hmean represents the crown height level of the "dominant layer"; m is obtained after normalization with Hmax, which is used to measure the crown layer uniformity and affect the horizontal continuity of crown fire.

[0210] Hmax: used as a reference height for scale normalization, making different stands comparable; also reflects the potential vertical channel length.

[0211] σh: crown height difference (layering degree); the greater the more layers and steps, the more conducive to upward penetration of fire.

[0212] C: canopy density, measures whether the horizontal combustible band is continuous; the greater the easier crown-to-crown propagation.

[0213] Hunder: understory shrub and grass layer thickness, directly corresponding to the abundance of "ladder fuel"; the greater the easier surface-to-crown.

[0214] Combustible species identification unit for extracting high VOC tree species coverage in each monitoring subunit, including: pine, eucalyptus, rhododendron, and laurel tree coverage, to build a volatile flammability index VEI for each monitoring subunit:

[0215]

[0216] wherein, represents the number of high VOC tree species, represents the volatile weight coefficient of the a-th tree species, wherein pine 0.8, eucalyptus 0.95, rhododendron 0.7, and laurel 0.6; represents the coverage of the a-th tree species in the r-th monitoring subunit;

[0217] The greater the VEI, the higher the proportion of high VOC tree species coverage in the unit, the higher the flammability, and the higher the potential fire risk and fire speed propagation possibility.

[0218] In this embodiment, the vegetation fuel characteristic analysis module realizes quantitative and refined evaluation of forest fire risk potential through fuel continuity analysis and combustible species identification. The fuel continuity analysis unit calculates the relative uniformity m of the canopy, the vertical stratification intensity s of the canopy, the relative amount u of the undergrowth "ladder fuel" thickness, and the horizontal connectivity c based on the average canopy height Hmean, the maximum canopy height Hmax, the height distribution standard deviation σh, the canopy density C, and the undergrowth shrub layer thickness Hunder of the monitoring subunit after clipping and normalization processing, which helps to quantify the lateral propagation potential of the canopy, the vertical fire potential connection ability, and the fire rising channel of the ground-undergrowth-canopy. The horizontal continuity index Hhor reflects the lateral continuity of the canopy, and the vertical continuity index Hver represents the potential of the ground fire leading to the canopy through the undergrowth combustible material. The combination of the two generates the fuel continuity index FCI. The greater the value, the more continuous the spatial fuel, the higher the fire propagation potential, which can scientifically reflect the contribution of the ladder fuel effect, the horizontal connectivity of the canopy, and the vertical stratification to the fire spread. The combustible species identification unit further extracts the coverage rate of high VOC tree species (such as pine, eucalyptus, rhododendron, and laurel) in each monitoring subunit, and combines with the weight coefficient of tree species volatile matter to construct the volatile flammable index VEI. The higher the value, the greater the proportion of high VOC tree species, the higher the flammability, and the greater the potential fire risk and fire speed propagation. Through this module, the vertical and horizontal fuel structure characteristics and combustible material distribution of the forest can be comprehensively and quantitatively characterized, which provides a scientific basis for fire risk comprehensive assessment and refined control strategy formulation, improves the accuracy and pertinence of fire risk early warning, and makes up for the neglect of vertical fuel continuity and high flammable tree species coverage in traditional methods.

[0219] Embodiment 8

[0220] This embodiment is an explanation and description in embodiment 1. Please refer to Figures 1 to 5 Specifically, the fire risk comprehensive assessment module includes a comprehensive fitting unit and a determination unit.

[0221] The comprehensive fitting unit is used to extract the wet-dry cycle index WD I, the monsoon effect index MFI, the dry thunderstorm risk factor DRF, the fuel continuity coefficient FCI, and the volatile flammable index VEI of each monitoring subunit. After normalization processing, the comprehensive fire risk index FFI of each monitoring subunit is constructed:

[0222]

[0223]

[0224] wherein, 、 、 、 、 is a weight;

[0225] a determination unit, configured to classify the risk level of the comprehensive fire danger index FFI of each monitoring subunit, specifically:

[0226] when FFI>0.7, the label is the first risk level, indicating that the monitoring subunit has continuous fuel, abundant combustible material, and meteorological conditions are not conducive to fire danger control, the risk of fire spread is high, and a first control strategy is generated, including: patrolling once a day, prohibiting all open fires, including barbecues, waste burning, and fireworks; setting obvious warning signs along the forest area and management roads, at least one "no open fire" sign every 500 meters; cleaning the 10-meter-wide strip area along the patrol route, forest roads, and forest edges within a 50-meter range of dry branches, fallen leaves, hay stacks, and waste wood at least once a week;

[0227] when 0.4≤FFI≤0.7, the label is the second risk level, indicating that the monitoring subunit has moderate fuel conditions and certain fire danger potential in meteorological conditions, and the fire spread speed is moderate: and a second control strategy is generated, including: patrolling once every two days, setting at least one "no open fire" sign every 800 meters; cleaning the 10-meter-wide strip area along the patrol route, forest roads, and forest edges within a 30-meter range of dry branches, fallen leaves, hay stacks, and waste wood once every two weeks;

[0228] when FFI<0.4, the label is the third risk level, indicating that the monitoring subunit has sparse or wet fuel and meteorological conditions conducive to suppressing fire spread, with low fire danger: and a third control strategy is generated, including: patrolling once a week, setting one "no open fire" sign every 1000 meters, and cleaning the 3-5-meter-wide strip area along the patrol route, forest roads, and forest edges within a 10-meter range of dry branches, fallen leaves, hay stacks, and waste wood once a month;

[0229] when the monitoring subunit is determined to be a 'high ignition probability area' and triggers the first early warning instruction, the monitoring subunits of the first early warning instruction are counted to form a first priority control group, and the corresponding control strategy is preferentially implemented.

[0230] In this embodiment, the comprehensive fitting unit and the risk level determination are combined to realize the quantitative evaluation and fine management of the fire risk potential of each monitoring subunit. The comprehensive fitting unit normalizes and weights the wet-dry cycle index WD I, the monsoon effect index MFI, the dry thunderstorm risk factor DRF, the fuel continuity coefficient FCI, and the volatile flammable index VEI to generate a comprehensive fire risk index FFI, which scientifically reflects the comprehensive influence of meteorological conditions, fuel structure, and combustible material distribution on fire spread. The determination unit divides the monitoring subunit into three risk levels according to the FFI value. The first risk level with FFI>0.7 indicates that the fuel is continuous, the combustible material is abundant, the meteorological conditions are not conducive to fire risk control, the fire spreads rapidly, and strict control strategies are triggered, such as high-frequency patrol, global fire ban, cleaning of dry branches and leaves along the patrol route and forest edge, etc. The second risk level with 0.4≤FFI≤0.7 indicates that the fuel condition is moderate, the meteorological conditions have certain fire risk potential, and the fire spread speed is moderate, triggering moderate-intensity control strategies such as reducing patrol frequency, increasing fire ban sign spacing, and moderate cleaning frequency. The third risk level with FFI<0.4 indicates that the fuel is sparse or wet, the meteorological conditions are conducive to suppressing fire spread, and the fire risk is low, triggering low-intensity control strategies such as extending patrol cycle, increasing sign spacing, and reducing cleaning frequency. At the same time, when the monitoring subunit is determined as a 'high ignition probability area' and triggers the first warning instruction, a first priority control group is formed, and corresponding control strategies are implemented to realize rapid response and optimal allocation of resources in high-risk areas. Through this module, dynamic and hierarchical fire risk warning and targeted control based on meteorological and fuel conditions can be realized, significantly improving the accuracy and effectiveness of forest fire risk prevention and control.

[0231] Embodiment 9

[0232] A forest fire risk warning method based on meteorological and vegetation data analysis, please refer to Figure 6 , comprising the following steps:

[0233] S1, obtain high-resolution digital elevation model DEM and multi-spectral or hyperspectral remote sensing image, divide the target forest area into regular grid according to the set spatial resolution, divide into several monitoring subunits, and identify and mark the vegetation type in each monitoring subunit;

[0234] S2, determine the monitoring date n and the latitude and longitude of the monitoring area based on global positioning system GPS and astronomical epoch calculation, divide the whole day into K time steps according to time step Δt, collect and obtain the hourly wind speed , wind direction , relative humidity , hourly precipitation and vapor pressure difference , a meteorological data set is established; a wet-dry cycle index WDI, a monsoon effect index MFI and a dry thunderstorm risk factor DRF are constructed, if the thunderstorm risk factor DRF = 1 at any moment, the monitoring subunit corresponding to the monitoring date is marked as a 'high ignition probability area', and a first early warning instruction is triggered, indicating that the monitoring subunit has a risk of lightning igniting dry vegetation or surface combustible materials, and through grid aggregation, a dry thunderstorm risk layer is output;

[0235] S3, extracting vegetation data in each monitoring subunit, calculating a fuel continuity coefficient FCI and a volatile flammable index VEI;

[0236] S4, based on the wet-dry cycle index WDI, the monsoon effect index MFI, the dry thunderstorm risk factor DRF, the fuel continuity coefficient FCI and the volatile flammable index VEI, a comprehensive fire danger index FFI of each monitoring subunit is constructed, and the comprehensive fire danger index FFI of each monitoring subunit is classified according to risk levels to obtain corresponding risk levels and corresponding control strategies, and combined with the first early warning instruction, a first priority control group is generated to preferentially implement the corresponding control strategies.

[0237] Figure 6The clearly visible grid represents the monitoring subunit for fine management of the target forest area, which corresponds to the method S1 step "regular grid division of the target forest area according to the set spatial resolution". The undergrowth shrubs, tree trunks and canopy layer structure, as well as the symbolic satellite, weather station and lightning symbols in the view represent the source of vegetation data and meteorological data acquisition, which corresponds to the process of "acquiring high-resolution digital elevation model DEM and multi-spectral or hyperspectral remote sensing image" in S1 step and the process of establishing meteorological data set in S2 step. The right side of the technical roadmap clearly reveals the logical processing flow of the method. The first box "data acquisition and grid division" summarizes S1 step; the second box "meteorological risk analysis" corresponds to the process of calculating WDI, MFI and DRF based on meteorological data in S2 step; the third box "combustible fuel feature quantification" corresponds to the process of calculating FCI and VEI based on vegetation data in S3 step; and the last box "comprehensive evaluation and graded early warning" corresponds to S4 step, that is, the final link of building a comprehensive fire danger index FFI based on all the above indexes, and risk grading and generating control strategies. In this embodiment, the target forest area is finely divided into grids by high-resolution DEM and multi-spectral or hyperspectral image, and the vegetation type of each monitoring subunit is identified to realize the acquisition of spatially accurate fuel distribution information; combined with hourly wind speed, wind direction, relative humidity, precipitation and vapor pressure difference, the wetting-drying cycle index WDI, the monsoon effect index MFI and the dry thunderstorm risk factor DRF are calculated, which can dynamically reflect the influence of extreme weather such as wind field mutation, heavy rain re-drying and dry thunderstorm on fire danger potential during the monsoon invasion or retreat; by extracting the canopy height, canopy density, undergrowth shrub layer thickness and high VOC tree coverage, the fuel continuity coefficient FCI and the volatile flammable index VEI are calculated to realize the accurate quantification of vertical ladder fuel and horizontal canopy continuity and flammability; the comprehensive fire danger index FFI is constructed by comprehensively considering the above meteorological and fuel indexes, and the monitoring subunit is graded according to the numerical value, and the first priority control group is formed by preferentially identifying the high ignition probability area, which realizes the dynamic, graded and targeted nature of fire danger warning, effectively improves the precision of forest fire danger prevention and control and the utilization efficiency of patrol resources, and significantly reduces the risk of fire occurrence and rapid spread.

[0238] The size of the threshold is set for easy comparison. The size of the threshold depends on the amount of sample data and the base number set by the person skilled in the art for each group of sample data; as long as it does not affect the proportional relationship of the parameters and the quantized values.

[0239] The above formulas are obtained by collecting a large amount of data for software simulation and selecting a formula close to the true value, and the coefficients in the formula are set by the person skilled in the art according to the actual situation. The above is only a preferred specific embodiment of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can make equivalent replacement or change according to the technical scheme and the inventive concept of the present application within the technical range disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A forest fire danger warning system based on weather and vegetation data analysis, characterized by, Comprise: A monitoring area division module, configured to divide a target forest area into a plurality of monitoring subunits according to a set spatial resolution based on a high-resolution digital elevation model (DEM) and remote sensing images, identify and extract a vegetation type in each monitoring subunit, and mark the vegetation type; A meteorological monsoon feature analysis module is used to determine the monitoring date n and the latitude and longitude of the monitoring area based on global positioning system (GPS) and astronomical epoch calculation, divide the whole day into K time steps Δt, collect the hourly wind speed , wind direction , relative humidity , hourly precipitation , and vapor pressure difference at the kth time index, and establish a meteorological data set to construct: A wet-dry cycle index (WDI) configured to represent a rapid fine fuel re-drying effect after heavy rain; A monsoon effect index (MFI) configured to represent an intensifying effect of a wind field mutation during a monsoon invasion or retreat period; And further identify dry thunderstorm characteristics, build dry thunderstorm risk factor DRF, when the hourly lightning frequency >5 times / hour, and the hourly precipitation <1mm / h, determine the high ignition probability area, and trigger the first warning instruction; A vegetation fuel feature analysis module configured to extract and calculate vegetation data in each monitoring subunit, including: A canopy height distribution, a canopy density, and an understory shrub and grass layer thickness, configured to calculate a fuel continuity index (FCI); An extractable flammable volatile matter ratio, including a pine, eucalyptus, rhododendron, and laurel coverage rate, configured to construct a volatile matter flammability index (VEI); A fire risk comprehensive assessment module configured to construct a comprehensive fire risk index (FFI) of each monitoring subunit based on the wet-dry cycle index (WDI), the monsoon effect index (MFI), a dry thunderstorm risk factor (DRF), the fuel continuity index (FCI), and the volatile matter flammability index (VEI), grade the comprehensive fire risk index (FFI) of each monitoring subunit, obtain a corresponding risk grade and a corresponding control strategy, and generate a first priority control group to implement the corresponding control strategy in priority based on the first early warning instruction.

2. The forest fire risk warning system based on weather and vegetation data analysis according to claim 1, characterized in that, The monitoring area division module comprises a data acquisition and preprocessing unit, a division unit, and a vegetation type identification unit; The data acquisition and preprocessing unit is configured to set an external data interface to receive or download multispectral or hyperspectral remote sensing image data from a satellite / airborne / unmanned aerial vehicle platform, and laser radar (LiDAR) point cloud and echo data from an airborne or ground platform, and perform radiation calibration, atmospheric correction, geometric correction, denoising, and cloud and snow masking processing; perform ground / non-ground classification, outlier removal, and intensity normalization on the LiDAR point cloud; unify the coordinate reference system and projection of the multispectral or hyperspectral remote sensing image data and the LiDAR point cloud and echo data, and perform spatio-temporal registration and co-location correction with the high-resolution digital elevation model (DEM), and resample and crop according to an initial spatial resolution R to generate a vegetation data set for subsequent analysis; The division unit is configured to divide the target forest area into a grid according to the initial spatial resolution R in the DEM coordinate system to obtain a monitoring subunit set and complete spatial labeling and indexing on the DEM; Monitoring subunit set , to denotes the 1st monitoring subunit to the Zth monitoring subunit; The vegetation type identification unit is configured to use a supervised classification method to classify the vegetation data set of the multispectral or hyperspectral image output by the data acquisition and preprocessing unit by using a support vector machine after training to output a vegetation category of each monitoring subunit, and the vegetation category includes a coniferous forest, a broad-leaved forest, a mixed forest, a shrub, and grassland.

3. The forest fire risk warning system based on weather and vegetation data analysis according to claim 2, characterized in that, The vegetation type identification unit comprises a feature extraction unit and a classification unit; The feature extraction unit is configured to extract spectral indexes based on the vegetation data set of the multispectral or hyperspectral image output by the data acquisition and preprocessing unit, the spectral indexes including a normalized difference vegetation index (NDVI), a red edge vegetation index (REVI), and a shortwave infrared moisture index (MSI), and the expressions are as follows: ； ； ； wherein NIR is near-infrared band reflectivity, RED is red band reflectivity, RE is red edge band reflectivity, and SWIR is shortwave infrared band reflectivity; A gray level co-occurrence matrix P(i,j) is constructed, and the specific steps are as follows: spatial texture features are extracted, specifically, the target band or principal component image is quantized to L=32 gray levels, the joint occurrence frequency of the gray level values of the neighborhood pixels of the reference pixel is obtained by using a window size w=7x7 pixels, a distance d=1, and a direction θ∈{0°, 45°, 90°, 135°} and the statistical results are normalized to obtain the gray level co-occurrence matrix P(i,j): ； wherein N(i,j) is the number of times of co-occurrence of the gray levels i and j; i represents the gray level value of the reference pixel, j represents the gray level value of the neighborhood pixel of the reference pixel in the specified direction and distance, and P(i,j) represents the joint occurrence probability of the pixel with the gray level i and the neighborhood pixel with the gray level j in the window; directional averages are calculated and obtained to obtain texture indexes, including: Contrast CON: ; Homogeneity HOM: ; Energy / Angle Second Moment ASM: ; Entropy ENT: ; wherein P(i,j) = 0 is ignored for the index entropy; COR: wherein, represents the row direction gray mean value, represents the column direction gray mean value, represents the row direction standard deviation, represents the column direction standard deviation, The classification unit is configured to perform classification based on the spectral indexes and the texture indexes, including: when the NDVI value of the monitoring subunit is between 0.65 and 0.85, the REVI value is between 0.10 and 0.20, the MSI value is between 0.45 and 0.55, the texture index contrast CON is between 10 and 20, the homogeneity HOM is between 0.70 and 0.85, the energy ASM is between 0.55 and 0.70, the entropy ENT is between 3 and 4, and the correlation COR is between 0.85 and 0.95, it is determined that the monitoring subunit is a coniferous forest; when the NDVI value of the monitoring subunit is between 0.60 and 0.80, the REVI value is between 0.15 and 0.25, the MSI value is between 0.50 and 0.60, the texture index CON is between 15 and 25, the HOM is between 0.60 and 0.75, the ASM is between 0.50 and 0.65, the ENT is between 3.5 and 4.5, and the COR is between 0.80 and 0.90, it is determined to be a broad-leaved forest; when the NDVI value of the monitoring subunit is between 0.55 and 0.75, the REVI value is between 0.12 and 0.22, the MSI value is between 0.50 and 0.65, the texture index CON is between 18 and 28, the HOM is between 0.55 and 0.70, the ASM is between 0.45 and 0.60, the ENT is between 4 and 5, and the COR is between 0.75 and 0.85, it is determined to be a mixed forest. When the NDVI value of the monitoring subunit is between 0.35 and 0.55, the REVI value is between 0.08 and 0.15, the MSI value is between 0.60 and 0.75, the texture index CON is between 22 and 35, the HOM is between 0.50 and 0.65, the ASM is between 0.40 and 0.55, the ENT is between 4.5 and 5.5, and the COR is between 0.70 and 0.80, it is determined as a shrub; When the NDVI value of the monitoring subunit is between 0.20 and 0.40, the REVI value is between 0.05 and 0.10, the MSI value is between 0.65 and 0.80, the texture index CON is between 12 and 20, the HOM is between 0.45 and 0.60, the ASM is between 0.35 and 0.50, the ENT is between 4 and 5, and the COR is between 0.65 and 0.75, it is determined as grassland.

4. The forest fire risk warning system based on weather and vegetation data analysis of claim 1, wherein, The meteorological monsoon feature analysis module includes a meteorological data acquisition unit, a wet-dry cycle index calculation unit, and a monsoon effect index calculation unit; The meteorological data acquisition unit is used to acquire hourly wind speed within the monitoring date n, using a time step Δt to form a time index k=1,…,K. ,wind direction relative humidity Hourly precipitation With vapor pressure difference Establish a meteorological dataset; and the scalar wind speed and wind direction are decomposed into a two-dimensional vector: , ; denotes the wind speed component in the east-west direction, denotes the wind speed component in the north-south direction; The wet-dry cycle index calculation unit is used to characterize the effect of strong precipitation rapidly wetting fine fuels, but subsequent sunny and high vapor pressure difference leading to rapid re-drying within 1-2 days, and is calculated using a recent W-hour window, with W=48h, calculating the wetness degree and the drying pull coefficient DI; ； ； wherein, W represents the latest W-hour set, is a scaling parameter, set to 10 mm, is the number of valid time points within the window, when the research object is evapotranspiration, is a valid time point; it indicates that the atmosphere has certain water deficit conditions, and the surface of the plant leaves begins to be affected by the evapotranspiration pull, determines the effective evapotranspiration period, and calculates which time points in the statistical window are valid samples; , which indicates that the air is extremely dry, exceeding the regulation capacity of the vegetation, and there is a risk of causing the vegetation stomata to close or even physiological stress; represents a reference vapor pressure difference risk identification threshold, set to 2 kPa, used to determine whether to enter the evapotranspiration stress state; Combined with the wetness and the dryness pull coefficient DI, dimensionless, the wet-dry cycle index WDI is calculated: 。 5. The forest fire risk warning system based on weather and vegetation data analysis according to claim 4, characterized in that, The monsoon effect index calculation unit is used to represent the role of wind field mutation and low-level jet intensification during the monsoon invasion / withdrawal period. When the wind speed / direction changes sharply, it is beneficial to fire spread and uphill movement. The wind vector difference is calculated to obtain the monsoon effect index MFI: ； wherein, represents a wind speed component in the east-west direction, represents a wind speed component in the north-south direction; and represents a change in the wind vector component between two adjacent times.

6. The forest fire danger early warning system based on weather and vegetation data analysis according to claim 5, characterized in that, The meteorological monsoon feature analysis module further comprises a dry thunderstorm risk identification unit, which is configured to identify the hourly lightning frequency and the hourly precipitation , and preset a lightning threshold and a dry thunderstorm precipitation threshold . The lightning threshold is set to 5 times / hour. The dry thunderstorm precipitation threshold is set to 1 mm / h. A thunderstorm risk factor DRF is defined: ； If the thunderstorm risk factor DRF is 1 at any time within the monitoring date n, the monitoring subunit corresponding to the monitoring date is marked as a high fire probability area, and a first warning instruction is triggered, indicating that the monitoring subunit has the risk of lightning igniting dry vegetation or surface combustible materials, and through grid aggregation, a dry thunderstorm risk layer is output.

7. The forest fire risk warning system based on weather and vegetation data analysis as claimed in claim 1, wherein, The vegetation fuel feature analysis module includes a canopy height distribution extraction unit, a canopy density calculation unit, and an understory shrub and grass layer thickness determination unit; The canopy height distribution extraction unit is used to extract the canopy height distribution curve within the monitoring subunit using LiDAR point cloud data or a three-dimensional point cloud model of the forest constructed from stereo image pairs, and calculate the average canopy height Hmean, the maximum canopy height Hmax, and the standard deviation σh of the height distribution: ； Wherein, N is the total number of point clouds in the monitoring subunit, is the height of the pth point; The canopy density calculation unit is used to calculate the canopy density C of the monitoring subunit based on the remote sensing image in the vegetation data set through point cloud porosity analysis technology, with a range of 0~1, representing the proportion of the ground covered by the canopy. The calculation formula of the canopy density C is: ； Where GapArea is the visible gap area of the ground, and TotalArea is the total area of the monitoring subunit; The understory shrub and grass layer thickness determination unit is used to calculate the average thickness Hunder of the shrub and herbaceous vegetation within the monitoring subunit based on the low-layer point cloud echo in the hyperspectral image of the vegetation data set, with the calculation formula being: ； wherein M is the number of low-level pixels within the monitoring subunit, Hq is the height of the qth low-level pixel, the low-level including heights between 0-2 m from the ground.

8. The forest fire risk early warning system based on weather and vegetation data analysis according to claim 7, characterized in that, The vegetation fuel feature analysis module further includes a fuel continuity analysis unit and a combustible species identification unit; The fuel continuity analysis unit is used for obtaining the relative uniformity m of the canopy, the vertical stratification intensity s of the canopy, the relative amount u of the ladder fuel thickness under the canopy, and the horizontal canopy density c by using a clipping function to normalize and boundary process the average canopy height Hmean, the maximum canopy height Hmax, and the standard deviation σh of the height distribution, the canopy density C, and the average thickness Hunder of shrubs and herbaceous vegetation; ； ； ； ； The clipping function is clip(x, 0, 1) = min(max(x, 0), 1); A horizontal continuity component is constructed to obtain a horizontal continuity index Hhor: ； A vertical continuity component is constructed to obtain a vertical continuity index Hver: ； wherein and are weights, , , the undergrowth thickness is slightly heavier than the degree of stratification; The fuel continuity coefficient FCI is calculated by combining the horizontal continuity index Hhor and the vertical continuity index Hver: ； wherein and are weights, , ; The combustible species identification unit is used to extract high VOC tree species coverage in each monitoring subunit, including pine, eucalyptus, rhododendron, and laurel tree coverage, and to construct a volatile flammability index VEI for each monitoring subunit: ； wherein, represents the number of high VOC species, represents the volatile weight coefficient of the a-th tree species, wherein, Pinus 0.8, Eucalyptus 0.95, Rhododendron 0.7, Lauraceae 0.6; represents the coverage rate of the a-th tree species in the r-th monitoring subunit.

9. The forest fire risk warning system based on weather and vegetation data analysis according to claim 1, characterized in that, The fire risk comprehensive assessment module includes a comprehensive fitting unit and a judgment unit; The comprehensive fitting unit is used to extract the wet-dry cycle index WD I, the monsoon effect index MFI, the dry thunderstorm risk factor DRF, the fuel continuity coefficient FCI, and the volatile flammability index VEI of each monitoring subunit, and to construct a comprehensive fire risk index FFI for each monitoring subunit after normalization processing: ； ； wherein , , , , is a weight; The judgment unit is used to classify the comprehensive fire risk index FFI of each monitoring subunit into risk levels, specifically: When FFI > 0.7, the label is the first risk level, and a first control strategy is generated, including: patrolling once a day, prohibiting all open fires, including barbecues, waste burning, and fireworks and firecrackers; and setting obvious warning signs along the forest area and management roads, at least one no-fire sign every 500 meters; cleaning the 10-meter-wide strip area along the patrol route, forest roads, and forest edges within a 50-meter range every week; When 0.4 ≤ FFI ≤ 0.7, the label is the second risk level, and a second control strategy is generated, including: patrolling once every two days, setting at least one no-fire sign every 800 meters; cleaning the 10-meter-wide strip area along the patrol route, forest roads, and forest edges within a 30-meter range every two weeks; When FFI < 0.4, the label is the third risk level, and a third control strategy is generated, including: patrolling once a week, setting one no-fire sign every 1000 meters, cleaning once a month, and cleaning the 10-meter-wide strip area along the patrol route, forest roads, and forest edges within a 10-meter range every week; When the monitoring subunit is determined to be a high ignition probability area and triggers the first early warning instruction, the monitoring subunits of the first early warning instruction are counted to form a first priority control group, and the corresponding control strategy is preferentially implemented.

10. A forest fire danger early warning method based on meteorological and vegetation data analysis, applied to a forest fire danger early warning system based on meteorological and vegetation data analysis according to any one of claims 1-9, characterized in that, The method includes the following steps: S1, acquire high-resolution digital elevation model DEM and multi-spectral or hyperspectral remote sensing image, divide the target forest area into several monitoring sub-units according to the set spatial resolution, and identify and mark the vegetation type in each monitoring sub-unit; S2, based on global positioning system GPS and astronomical epoch calculation to determine the monitoring date n and the latitude and longitude of the monitoring area, dividing the whole day into K time steps Δt, collecting the hourly wind speed of the kth time index , wind direction , relative humidity , hourly precipitation and vapor pressure difference , and establishing a meteorological data set; The wet-dry cycle index WDI, the monsoon effect index MFI and the dry thunderstorm risk factor DRF are constructed. If the thunderstorm risk factor DRF = 1 at any time, the monitoring sub-unit corresponding to the monitoring date is marked as a high ignition probability area, and a first warning instruction is triggered, indicating that the monitoring sub-unit has the risk of lightning igniting dry vegetation or surface combustible material, and through grid aggregation, a dry thunderstorm risk layer is output; S3, extract the vegetation data in each monitoring sub-unit, calculate and obtain the fuel continuity coefficient FCI and the volatile flammable index VEI; S4, based on the wet-dry cycle index WDI, the monsoon effect index MFI, the dry thunderstorm risk factor DRF, the fuel continuity coefficient FCI and the volatile flammable index VEI, the comprehensive fire danger index FFI of each monitoring sub-unit is constructed, and the comprehensive fire danger index FFI of each monitoring sub-unit is classified according to the risk level, the corresponding risk level and the corresponding control strategy are obtained, and the first priority control group is generated by combining the first warning instruction to implement the corresponding control strategy.

Images