Forest fuel ecological regulation effect evaluation system based on multi-modal data

Abstract

The present application relates to the technical field of forest fire prevention management, and discloses a forest combustible ecological regulation effect evaluation system based on multi-modal data, which comprises a multi-modal acquisition module, a regional load evaluation module, a regional pairing module, a combustible risk evaluation module and an ecological regulation module. The system obtains forest terrain, vegetation management and environmental sensing data through the multi-modal acquisition module, comprehensively evaluates the influence degree of the superposition effect on combustible accumulation, retention and flammability through the regional load evaluation module, provides clear spatial basis for hierarchical regulation, dynamically constructs the pairing relationship between the fire source origin and the leeward neighborhood based on the wind direction through the regional pairing module, locks the potential spread path of the fire, evaluates the combustible development trend within the range of the regional pairing group through the combustible risk evaluation module, judges the load grade of each region of the forest and the fire danger grade of the regional pairing group through the ecological regulation module, triggers the corresponding ecological regulation response measures, and realizes long-acting optimization of combustible and closed-loop control of fire risk.

Description

Technical Field

[0001] This invention relates to the field of forest fire prevention and management technology, specifically to a forest combustible ecological regulation effect evaluation system based on multimodal data. Background Technology

[0002] Forest combustibles refer to all organic matter in a forest ecosystem that, under specific conditions, can absorb heat, ignite, and sustain the spread of fire. Combustible load is one of the core material bases and key variables influencing the probability of fire occurrence, spread rate, combustion intensity, and the degree of ecological and economic losses caused by fire. Forest combustibles can be divided into aboveground and underground combustibles based on spatial distribution and combustion characteristics. Aboveground combustibles are the core monitoring targets, including dead branches, fallen branches, and decaying wood from the tree layer; whole plants from the shrub layer; dead and living herbs from the herbaceous layer; and surface layers of dead leaves, moss, and lichen. Underground combustibles mainly consist of humus and dead roots. Although they are more difficult to ignite, they burn for a longer period, easily forming underground fires and increasing the difficulty of firefighting. Accurately understanding the forest combustible load is a prerequisite for scientifically delineating fire risk levels, formulating forest fire prevention plans, and carrying out ecological regulation of combustibles. It is also a key foundation for improving the scientific nature and effectiveness of forest fire prevention and control.

[0003] Currently, existing forest combustible ecological regulation methods mainly focus on reducing combustible load, optimizing combustible structure, regulating the forest ecological environment, and reducing forest fire risk. Ecological regulation primarily utilizes forest ecological processes to inhibit the continuous accumulation of combustibles, with common methods including adjusting stand density, introducing fire-resistant tree species, and optimizing the tree, shrub, and grass vegetation structure. However, traditional forest combustible ecological regulation effectiveness assessment systems, in practical applications, tend to overlook the impact of micro-topographical changes such as altitude and slope on the spatial distribution of combustibles. Furthermore, they fail to fully integrate the coupling effects between topographic, vegetation, and environmental conditions, leading to insufficient reliability in evaluating the effectiveness of combustible ecological regulation. In addition, existing systems typically do not adequately consider the forest fire risk propagation paths and regional risk transmission characteristics driven by wind direction, making it difficult to identify high-risk diffusion areas in advance and plan targeted firebreaks. This results in delayed forest fire prevention measures and a lack of effective proactive early warning and forward-looking planning capabilities. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a forest combustible ecological regulation effect evaluation system based on multimodal data. It has advantages such as multidimensional locking of potential fire spread paths and good long-term optimization and regulation effects, and solves the problems of insufficient reliability of traditional forest combustible ecological regulation effect evaluation systems and lack of forward-looking planning capabilities for fire risk control measures.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a forest combustible ecological regulation effect evaluation system based on multimodal data, comprising: The multimodal acquisition module is used to acquire topographic data, vegetation management data, and environmental sensing data from all areas of the forest, and classify them into forest datasets, vegetation datasets, and environmental datasets. The regional load assessment module is used to comprehensively evaluate the combined effects of topography, vegetation, and environmental conditions in different forest areas on the accumulation, retention, and flammability of combustibles, and outputs the corresponding load index. ; The region matching module has a sliding window with a fixed duration. Generate corresponding area pairing groups according to wind direction; The combustible risk assessment module is used to assess the combustible development trend within the paired area and output the corresponding risk index. ; The ecological regulation module is set with a fixed range of thresholds to determine the load level of each forest area and the fire risk level of the paired regional groups, output the assessment results and trigger corresponding ecological regulation response measures.

[0006] Preferably, each area of ​​the forest has several monitoring points, and the forest dataset includes the total number of monitoring points, altitude, slope, and aspect angle for each area.

[0007] Preferably, the vegetation dataset includes aboveground biomass, dead combustible load, humus layer thickness, humus layer moisture content, moss and lichen layer thickness, moss and lichen layer moisture content, herbaceous layer cover, live combustible load, shrub layer cover, shrub layer dead branches, tree layer canopy closure, and tree layer dead branches for each region.

[0008] Preferably, the environmental dataset includes the average daily sunshine duration, thermal radiation intensity, surface temperature, canopy temperature, and wind speed for each region.

[0009] Preferably, the load index The calculation process is as follows: S11. Based on the forest dataset, vegetation dataset, and environment dataset, extract the first... Topographic data, vegetation management data, and environmental sensor data of the region at the current timestamp; S12, Calculate the first Average altitude of the region at the current timestamp and standard deviation ; S13, Calculate the... Average slope of the current timestamp in the region and standard deviation ; S14. Calculate the weighted average. Reserve index of the region at the current timestamp ; S15, Determine the first The slope aspect type of each monitoring point in the area was determined, and the first... The proportion of shady slopes in the area The proportion of sunny slopes and the proportion of semi-sunny slopes ; S16. Calculate the weighted average. Decomposition adjustment index of the current timestamp of the region ; S17, Calculate the... Average thickness of humus layer at the current timestamp of the region Average moisture content of humus layer Mean thickness of moss and lichen layer Mean moisture content of moss and lichen layers and surface humus index ; S18. Calculate the first using a weighted method. Distribution index of current timestamp in the region ; S19. Based on S11-S18, calculate the first... Load index of the region at the current timestamp .

[0010] Preferably, the region pairing process is as follows: S21. Extract the sliding window based on the environment dataset. Within, environmental sensor data for each area of ​​the forest; S22, Calculate the sliding window Relative range of thermal radiation intensity at the start timestamp ; S23, Based on the sliding window Relative range of thermal radiation intensity at the start timestamp Determine the origin of the forest area The process is as follows: If sliding window Relative range of thermal radiation intensity at the start timestamp This indicates that the thermal radiation intensity is completely equal across all areas of the forest, and the geometric center of the forest area is selected as the origin of the forest region. ; If sliding window Relative range of thermal radiation intensity at the start timestamp This indicates that there are differences in the intensity of thermal radiation in different areas of the forest. The area with the maximum intensity of thermal radiation in the forest is selected as the origin of the forest area. If multiple regions with the highest thermal radiation intensity exist, select the region with the shortest straight-line distance to the geometric center of the forest area as the origin of the forest region. ; S24. Determine the wind direction type for each timestamp and establish the origin for the forest area. The pairing process is as follows: The origin of the forest area The adjacent area to due north is denoted as The origin of the forest area The adjacent area to the northeast is denoted as The origin of the forest area The adjacent area to the due east direction is denoted as The origin of the forest area The adjacent area to the southeast is denoted as The origin of the forest area The adjacent area to the due south direction is denoted as The origin of the forest area The adjacent area to the southwest is denoted as The origin of the forest area The adjacent area to the due west direction is denoted as The origin of the forest area The adjacent area to the northwest is denoted as ; If 0° ≤ sliding window Wind direction at start timestamp ≤22.5°, determined to be a northerly wind, the origin point of the forest area is set. Adjacent area to due south Forming pairs; If 22.5° < sliding window Wind direction at start timestamp <67.5°, determined to be a northeasterly wind, the origin point of the forest area. Adjacent area to the southwest Forming pairs; If 67.5° ≤ sliding window Wind direction at start timestamp ≤112.5°, determined to be easterly, the origin point of the forest area is set. Adjacent area to due west Forming pairs; If 112.5° < sliding window Wind direction at start timestamp <157.5°, determined to be a southeast wind, the origin point of the forest area is set. Adjacent area to the northwest Forming pairs; If 157.5° ≤ sliding window Wind direction at start timestamp ≤202.5°, determined to be a southerly wind, the origin point of the forest area is set. Adjacent area to due north Forming pairs; If 202.5° < sliding window Wind direction at start timestamp <247.5°, determined to be a southwest wind, the origin point of the forest area is set. Adjacent area to the northeast Forming pairs; If 247.5° ≤ sliding window Wind direction at start timestamp ≤292.5°, determined to be westerly wind, the origin point of the forest area is set. Adjacent area to due east Forming pairs; If 292.5° < sliding window Wind direction at start timestamp <337.5°, determined to be a northwest wind, the origin point of the forest area is set. Adjacent area to the southeast Forming pairs; If 337.5° ≤ sliding window Wind direction at start timestamp ≤360°, determined to be a northerly wind, the origin of the forest area is set. Adjacent area to due south Form pairs.

[0011] Preferably, the risk index The calculation process is as follows: S31, Given a sliding window Within this dataset, all timestamps show a southerly wind. Then, based on the environmental dataset, the origin point for the forest region is extracted. Adjacent area to due north Environmental sensing data; S32, Calculate the sliding window Within, the origin of the forest area Adjacent area to due north Mean of absolute difference in load index ; S33, Calculate the sliding window Within, the origin of the forest area Surface temperature change rate Forest area origin Canopy temperature change rate Adjacent area to the due north direction Surface temperature change rate Adjacent area to the due north direction Canopy temperature change rate and the origin of the forest area Adjacent area to due north absolute difference in the rate of change of surface temperature absolute difference between the rate of change of canopy temperature and ; S34, Calculate the sliding window Within, the origin of the forest area average wind speed Adjacent area to the due north direction average wind speed and the origin of the forest area Adjacent area to due north absolute difference in average wind speed ; S35. Based on S31-S34, calculate the sliding window using a weighted method. Within, the origin of the forest area Adjacent area to due north Risk index of paired groups .

[0012] Preferably, the load level assessment process is as follows: The upper limit of the load threshold range is denoted as The lower limit of the load threshold range is denoted as ; If the first Regional load index < , indicating the first The area has low combustible material accumulation, is difficult to retain, and has weak flammability; its load level is Level 1. Response measures include maintaining the patrol frequency of the current assessment cycle, increasing the patrol frequency of the next assessment cycle, and implementing routine combustible material monitoring. No additional investment in thermal infrared imaging monitoring equipment is required. ≤th Regional load index ≤ , indicating the first The area has a moderate accumulation of combustibles, moderate retention difficulty, and moderate flammability, with a load level of 2. Response measures include increasing the frequency of patrols and combustible cleanup at the current timestamp, prioritizing manual and mechanical brush cutting and pruning to reduce surface and canopy combustible loads, and investing additional resources in thermal infrared imaging monitoring equipment for dynamic monitoring of key areas. If the [unclear text - possibly a missing section]... Regional load index > , indicating the first The area has a high accumulation of combustibles, is easy to retain, and is highly flammable, with a load level of 3. Response measures include immediately carrying out comprehensive combustible cleanup, developing a burning plan to optimize forest stand density and vegetation structure within the area, and investing additional resources in thermal infrared imaging monitoring equipment to implement daily full-coverage patrols until the [date missing]. Regional load index < .

[0013] Preferably, the fire risk assessment process is as follows: Let the upper limit of the risk threshold range be denoted as The lower limit of the risk threshold range is denoted as ; For sliding windows Within, if the origin of the forest area Adjacent area to due north Risk index of paired groups < This indicates that there is no obvious flammability trend within the area of ​​the paired group, the fire hazard level is 1, and the response measures include continuing the current sliding window. The patrol frequency is sufficient, eliminating the need for additional firebreaks. ≤ Forest region origin Adjacent area to due north Risk index of paired groups ≤ This indicates that there is a flammable tendency within the paired area, with a fire hazard level of 2. Response measures include increasing the sliding window. Increase patrol frequency within the area, activate preventative combustible material control, prepare fire extinguishing equipment, and establish additional firebreaks. (If the forest area is at its origin...) Adjacent area to due north Risk index of paired groups > This indicates a significant flammability trend within the paired area, with a fire risk level of 3. Response measures include real-time, full-coverage patrols, increasing the deployment of fire extinguishers and firebreaks, timely issuance of fire warnings, and prohibiting all open flame operations within the area, extending to the forest's origin point. Adjacent area to due north Risk index of paired groups < .

[0014] Preferably, the ecological regulation module targets the entire forest area. When the load level of three adjacent areas all reaches level 2 or above, or the fire risk level all reaches level 2 or above, comprehensive ecological regulation response measures are activated. These measures include promptly notifying management personnel to create firebreaks with a width of ≥10 meters and a depth of ≥30 centimeters. The surface of the firebreaks is cleaned twice a week. Combined with planned burning and mechanical cleaning measures, the combustible load in the firebreaks and surrounding areas is reduced. The surface of the firebreaks is cleaned twice a week. After cleaning, the regional load index needs to be assessed. and risk index If the results of two consecutive assessments drop below level 2, the frequency of cleanup can be appropriately reduced; otherwise, the existing comprehensive ecological regulation and response measures should continue to be implemented.

[0015] Compared with existing technologies, this invention provides a forest combustible ecological regulation effect evaluation system based on multimodal data, which has the following beneficial effects: 1. This invention employs a multimodal acquisition module with a 20m×20m grid partitioning system, deploying monitoring points according to a uniform distribution principle. The monitoring density is flexibly adjusted for complex terrain, effectively reducing sampling errors and accurately reflecting the true distribution of combustibles. It comprehensively acquires forest topography, vegetation management, and environmental sensor data, classifying and constructing standardized datasets to provide accurate information for assessing the ecological regulation effect of combustibles. The regional load assessment module integrates the superimposed effects of topography, vegetation, and environmental conditions to quantify combustible accumulation and flammability, outputting the corresponding load index. This provides a clear spatial basis for tiered regulation. The regional pairing module dynamically constructs pairing relationships between the origin of the fire source and its downwind neighboring area based on wind direction, accurately locating the regional origin. It also completed the pairing of adjacent areas to lock down potential fire spread paths.

[0016] 2. This invention assesses the flammability development trend within a paired area using a flammability risk assessment module, and outputs the corresponding risk index. This enables an upgrade from static load assessment to dynamic risk prediction. The ecological regulation module has a fixed range of thresholds to determine the load level of each forest area and the fire risk level of the paired area groups. It outputs the assessment results and triggers corresponding ecological regulation response measures. Through active intervention measures such as cleanup, burning, and isolation, it achieves long-term optimization of combustibles and closed-loop management of fire risks. Attached Figure Description

[0017] Figure 1 This is a system flowchart of the present invention. Detailed Implementation

[0018] 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.

[0019] Example Please see Figure 1 Table 1 shows the experimental data of the load index, Table 2 shows the experimental data of the paired regional groups, and Table 3 shows the experimental data of the risk index. This invention provides a forest combustible ecological regulation effect evaluation system based on multimodal data, including a multimodal acquisition module, a regional load assessment module, a regional pairing module, a combustible risk assessment module, and an ecological regulation module. The multimodal acquisition module is used to acquire topographic data, vegetation management data, and environmental sensing data from all areas of the forest, and classify them into forest datasets, vegetation datasets, and environmental datasets. Several monitoring points were set up in each area of ​​the forest. The forest dataset includes the total number of monitoring points, altitude, slope and aspect angle for each area. Specifically, forest grid management typically uses a 20m×20m scale to divide areas. This operational accuracy can effectively reflect the true distribution of combustibles within the area. Monitoring points in each area are set up according to the principle of uniform distribution. For areas with complex terrain, sampling errors can be reduced by appropriately increasing the number of monitoring points and reducing the spacing between them. The vegetation dataset includes aboveground biomass, dead combustible load, humus layer thickness, humus layer moisture content, moss and lichen layer thickness, moss and lichen layer moisture content, herbaceous layer cover, live combustible load, shrub layer cover, shrub layer dead branches, tree layer canopy closure, and tree layer dead branches for each region. The environmental dataset includes the average daily sunshine duration, thermal radiation intensity, surface temperature, canopy temperature, and wind speed for each region; The regional load assessment module is used to comprehensively evaluate the combined effects of topography, vegetation, and environmental conditions in different forest areas on the accumulation, retention, and flammability of combustibles, and outputs the corresponding load index. The calculation process is as follows: S11. Based on the forest dataset, vegetation dataset, and environment dataset, extract the first... Topographic data, vegetation management data, and environmental sensor data of the region at the current timestamp; S12. For the current timestamp, the first... The altitude of each monitoring point in the area is recorded as follows: , Indicates the first The total number of monitoring points in the area, then calculate the number of monitoring points. Average altitude of the region at the current timestamp and standard deviation Its expression is as follows: In the formula, Indicates the first Region 1 The altitude of each monitoring point ; Specifically, average altitude Higher altitudes typically result in lower regional temperatures and increased humidity, which inhibits the decomposition of combustibles and promotes increased litter accumulation. However, high altitudes also lead to reduced vegetation biomass and decreased live combustible resources. Therefore, a correlation analysis based on surface biomass and altitude standard deviation is necessary. The calculation significance is to quantify the degree of elevation variation of micro-topography in the region, thereby reflecting and explaining the spatial heterogeneity of combustible load; S13. For the current timestamp, the first... The slope of each monitoring point in the area is recorded as follows: Then calculate the first Average slope of the current timestamp in the region and standard deviation Its expression is as follows: In the formula, Indicates the first Region 1 The slope of each monitoring point; Specifically, average slope The larger the slope, the stronger the surface runoff and erosion, the more difficult it is for litter to accumulate, and the lower the surface combustible load. Since the vegetation on steep slopes is mostly drought-resistant shrubs, its flammability is significantly increased. Therefore, a correlation analysis needs to be conducted in conjunction with the dead combustible load, and the slope standard deviation should be considered. The larger the area, the higher the spatial heterogeneity of the forest area's combustible load, and the load of different types of combustibles will show differentiated distribution characteristics; S14. For the current timestamp, the first... The aboveground biomass of the region is denoted as , will the The dead combustible load of the area is recorded as Then calculate the weighted average. Reserve index of the region at the current timestamp Its expression is as follows: In the formula, This represents a reference value indicating the average altitude. Reference values ​​representing the standard deviation of altitude. This represents a reference value indicating the average slope. Reference values ​​representing the standard deviation of slope. Reference values ​​representing aboveground biomass. This indicates a reference value for the load of dead combustible materials. , , , , and All are weighting coefficients, and ; Specifically, the reference values ​​are the average values ​​of typical historical years in the region, or the benchmark values ​​normalized based on historical extreme values. After standardizing each indicator to eliminate dimensional differences, weights are assigned according to the degree of influence of each indicator on combustible material assessment. This not only quantifies the overall topographic level and micro-topographic undulation heterogeneity of the region, but also links the basic characteristics of the living and dead combustion carriers of combustible materials. It can comprehensively reflect the influence of topography on the distribution, accumulation and flammability of combustible materials, providing a scientific quantitative basis for the delineation of key monitoring areas and effectively improving the comprehensiveness and pertinence of combustible material assessment in forest grid management. S15. For the current timestamp, the first... The slope angle of each monitoring point in the area is denoted as . Then judge the first The slope aspect type for each monitoring point in the area is determined as follows: The first Region 1 The slope angle of each monitoring point is denoted as... , ; If 0°≤ <45° or 315°≤ ≤360°, classified as a shady slope; If 45°≤ <135° or 225°≤ ≤315°, classified as a semi-sunny slope; If 135°≤ <225°, classified as a sunny slope; According to statistics, the first The proportion of shady slopes in the area is denoted as , will the The percentage of sunny slopes in the area is denoted as , will the The proportion of semi-sunny slopes in the area is denoted as ; Specifically, the north-facing slope angle is 0°, the east-facing slope angle is 90°, the south-facing slope angle is 180°, and the west-facing slope angle is 270°. In the above classification, the shady slope (northward) and the sunny slope (southward) each cover 90°, and the semi-sunny slope (eastward and westward) covers a total of 180°, which is consistent with the gradual influence of slope aspect on sunlight and temperature and humidity in the Northern Hemisphere. S16. For the current timestamp, the first... The average daily sunshine duration of the region is recorded as Then calculate the weighted average. Decomposition adjustment index of the current timestamp of the region Its expression is as follows: In the formula, This indicates the weight of the group associated with the percentage of shady slopes and the degree of lack of sunlight. This represents a reference value indicating the average daily sunshine duration. This indicates the weight of the group associated with the percentage of sunny slopes and sufficient sunlight. This indicates the weight of the proportion of the half-sunny slope. , and All are constants, and ; Specifically, decompose the adjustment index Focusing on the correlation between "slope aspect - environmental factors - combustible characteristics," a higher index value indicates that the overall environment of the area is less conducive to the decomposition of combustibles and has a stronger tendency for their accumulation. Shady slopes have less sunlight and higher humidity, which inhibits the decomposition of dead combustibles and promotes their accumulation. Therefore, the proportion of shady slopes and the degree of sunlight deficiency are used as indicators. A positive correlation is formed between sunny slopes and sufficient sunlight and high temperatures, which promotes the decomposition of dead combustibles and reduces their accumulation. Therefore, the proportion of sunny slopes and the degree of sunlight are considered as positive correlations. The negative correlation group is formed, with the light and humidity of the semi-sunny slope falling between the two, and it is added to the comprehensive evaluation as an independent neutral moderating term; S17. For the current timestamp, the first... The thickness of the humus layer at each monitoring point in the area is recorded as follows: , will the The moisture content of the humus layer at each monitoring point in the area is recorded as follows: , will the The thickness of the moss and lichen layer at each monitoring point in the area is recorded as follows: , will the The moisture content of the moss and lichen layer at each monitoring point in the area was recorded as follows: Then calculate the first Average thickness of humus layer at the current timestamp of the region Average moisture content of humus layer Mean thickness of moss and lichen layer Mean moisture content of moss and lichen layers and surface humus index Its expression is as follows: In the formula, Indicates the first Region 1 Thickness of humus layer at each monitoring point Indicates the first Region 1 Moisture content of the humus layer at each monitoring point; In the formula, Indicates the first Region 1 The thickness of the moss and lichen layer at each monitoring point Indicates the first Region 1 Moisture content of moss and lichen layers at each monitoring point; In the formula, This represents a reference value indicating the average thickness of the humus layer. This represents a reference value indicating the average moisture content of the humus layer. This represents a reference value indicating the average thickness of the moss and lichen layer. This represents a reference value indicating the average moisture content of the moss and lichen layer. Indicate the stabilizing factor to avoid a denominator of 0; Specifically, the humus layer and the moss and lichen layer are both surface combustibles. The thickness directly determines the combustible load level, while the moisture content is the core factor affecting the risk of smoldering and flammability. The thickness combined with the moisture content can more accurately quantify the key characteristics of forest fire risk. The greater the thickness and the lower the moisture content, the higher the effective combustible contribution, thus more accurately quantifying the potential combustion contribution of surface combustibles. S18. For the current timestamp, the first... The herbaceous cover of the region is denoted as , will the The active combustible load of the area is denoted as , will the The shrub cover of the area is denoted as , will the The amount of dead branches in the shrub layer of the area is recorded as follows: , will the The canopy closure of the tree layer in the region is denoted as , will the The amount of dead branches in the tree layer of the area is recorded as follows: Then calculate the weighted average. Distribution index of current timestamp in the region Its expression is as follows: In the formula, Reference values ​​indicating herbaceous cover. Indicates the first The ratio of dead to fresh in the region This indicates the weight of the group relating herbaceous cover and the ratio of withered to fresh herbs. Reference values ​​indicating shrub cover. Reference values ​​for the amount of dead branches in the shrub layer. This indicates the weight of the association group between shrub cover and shrub dead branches. Reference values ​​indicating the canopy closure of the tree layer. Reference values ​​for the amount of dead branches in the tree layer. This indicates the weight of the association group between canopy closure and dead branches in the canopy. , and All are constants, and ; Specifically, through the distribution index The core parameters of the three layers of combustibles in forests—herbaceous plants, shrubs, and trees—are linked, and the dimensions of the product terms for each layer are unified as "standardized coverage × standardized load," which comprehensively quantifies the spatial distribution and flammability characteristics of combustibles in each layer and improves the pertinence of subsequent assessments. S19. Based on S11-S18, calculate the first... Load index of the region at the current timestamp Its expression is as follows: In the formula, , , and All are weights, and ; Specifically, load index The risk of regional combustibles is comprehensively quantified from four dimensions: reserves, decomposition and regulation, surface humus layer, and vertical distribution. Among them, the reserves index... Decomposition Regulation Index (DRI) is used to quantify the static accumulation potential of combustibles under the combined effects of topography and biomass. The surface humus index is used to quantify the dynamic environmental regulation of decomposition rate by slope aspect and light intensity (higher values ​​indicate a stronger tendency to accumulate). Distribution index used to quantify the effective combustible contribution of humus and moss-lichen layers. It is used to quantify the effective load and spatial distribution characteristics of each vertical level from herbaceous to shrub to tree, and comprehensively covers the four key dimensions of combustible load: "stock base - environmental trend - surface details - vertical structure", providing a scientific quantitative basis for the delineation of key monitoring areas and forest grid management; The following are the experimental data for the load index, as shown in Table 1: Table 1: Experimental Data of Load Index In Table 1, the experimental data of the load index were used to select area A, which completed a single cycle of ecological regulation in the forest, as the experimental object. Five monitoring points were set up in area A. Reserves Index The weights are set as follows: , , , , , ; Decomposition adjustment index The weights are set as follows: , , ; Surface humus index In the middle, the stability factor ; Distribution index The weights are set as follows: , , ; Load index The weights are set as follows: , , , ; The ecological regulation module is set with a fixed range of load thresholds. This method is used to quickly determine the load level of different areas of a forest, providing a quantitative reference for the formulation of forest fire prevention and control strategies. The rationality of the range directly affects the accuracy of the load level assessment and the timeliness of prevention and control interventions. An excessively narrow load threshold range... This could lead to overly strict load level classification, misclassifying a load level 1 area as a level 2 or 3, increasing unnecessary combustible material cleanup costs and wasting equipment resources, while also increasing the workload of prevention and control personnel. Conversely, overly strict load level classification would blur the lines between level 1 and level 2, and level 2 and level 3 fire risk, delaying targeted prevention and control in high-risk areas and creating significant forest fire hazards. Therefore, the optimal range of this threshold interval needs to be determined through the following system calibration experiments: Load Threshold Interval The calibration method is as follows: Forest fire prevention databases were used to screen forest area samples at different fire load levels, covering low, medium, and high combustible material accumulation, as well as various terrain and vegetation types. Combustible material monitoring data (such as thickness of fallen leaves and branches, vegetation coverage, and moisture content), historical fire records, and prevention and control effectiveness data (such as fire probability, spread rate, and control effectiveness) were extracted from the sample areas. Different candidate threshold ranges were set, and in each calibration experiment, the fire load level of the sample areas was classified based on the candidate threshold range. The matching degree between the classification results and the gold standard for actual fire risk assessment was recorded. Then, combined with dynamic risk monitoring data, the regional fire load index under different prevention and control intervention intensities was simulated. The changing trend was analyzed, and the upper and lower limits of the candidate threshold intervals were adjusted. Multiple verification experiments were conducted to record the impact of threshold interval settings on the evaluation of prevention and control effectiveness. For each candidate threshold interval, the collected sample data and dynamic simulation results were used as inputs to count the number of times low-level loads were misjudged as high-level loads due to improper interval settings (counted as over-evaluation), the number of times high-level loads were misjudged as low-level loads (counted as under-evaluation), and the degree of fit between the load level classification results and the actual prevention and control effectiveness (such as fire incidence rate, prevention and control resource utilization rate, etc.). Finally, the interval range that minimizes both the over-evaluation rate and the under-evaluation rate and has the highest degree of fit with the actual prevention and control effectiveness was selected as the load threshold interval. The preferred range; Table 1 shows the load threshold range in the load index experimental data. The preferred range is defined as 0.6-0.9. It is determined that 0.6 < the load index of region A. <0.9 indicates that the amount of combustible material accumulated in forest area A is moderate, the retention difficulty is moderate, and the flammability is moderate. The load level is 2. Response measures include increasing the frequency of patrols and combustible material cleanup at the current timestamp, prioritizing manual and mechanical brush clearing and pruning to reduce the amount of combustible material on the ground and in the canopy, and investing additional resources in thermal infrared imaging monitoring equipment to implement dynamic monitoring of key areas. The region matching module has a sliding window with a fixed duration. The corresponding regional pairing groups are generated according to the wind direction. The regional pairing process is as follows: S21. Extract the sliding window based on the environment dataset. Inside, environmental sensor data for each area of ​​the forest is used to create a sliding window. The total number of timestamps within is denoted as The total number of forest areas is denoted as ; S22, For sliding windows The starting timestamp records the thermal radiation intensity of each area of ​​the forest as follows: Then calculate the sliding window. Relative range of thermal radiation intensity at the start timestamp Its expression is as follows: In the formula, Represents a sliding window Start timestamp, forest number The thermal radiation intensity of each region , Represents a sliding window Start timestamp, mean thermal radiation intensity of the forest This indicates the maximum intensity of thermal radiation in the forest. This represents the minimum intensity of thermal radiation in a forest. S23, Based on the sliding window Relative range of thermal radiation intensity at the start timestamp Determine the origin of the forest area The process is as follows: If sliding window Relative range of thermal radiation intensity at the start timestamp This indicates that the thermal radiation intensity is completely equal across all areas of the forest, and the geometric center of the forest area is selected as the origin of the forest region. ; If sliding window Relative range of thermal radiation intensity at the start timestamp This indicates that there are differences in the intensity of thermal radiation in different areas of the forest. The area with the maximum intensity of thermal radiation in the forest is selected as the origin of the forest area. If multiple regions with the highest thermal radiation intensity exist, select the region with the shortest straight-line distance to the geometric center of the forest area as the origin of the forest region. ; Specifically, the relative range of thermal radiation intensity The smaller the value, the more uniform the thermal radiation intensity across regions, the smaller the differences between regions, the more concentrated the data distribution, the lower the dispersion, and the greater the relative range of thermal radiation intensity. The larger the value, the greater the difference in thermal radiation intensity in different areas, the more dispersed the data distribution, the higher the degree of dispersion, and the more prominent the local extreme value characteristics. It usually corresponds to the occurrence of local high temperature points, thermal anomalies, fire hazards, and areas with concentrated heat sources. It is a typical high fire risk and suspected fire anomaly characteristic signal in fire monitoring. S24. Move the sliding window Within, the wind direction angle for each timestamp is recorded as... , Represents a sliding window The wind direction angle at the starting timestamp is used to determine the wind type for each timestamp and to establish the origin for the forest region. The pairing process is as follows: The origin of the forest area The adjacent area to due north is denoted as The origin of the forest area The adjacent area to the northeast is denoted as The origin of the forest area The adjacent area to the due east direction is denoted as The origin of the forest area The adjacent area to the southeast is denoted as The origin of the forest area The adjacent area to the due south direction is denoted as The origin of the forest area The adjacent area to the southwest is denoted as The origin of the forest area The adjacent area to the due west direction is denoted as The origin of the forest area The adjacent area to the northwest is denoted as ; Based on the physical laws of fire: wind direction is defined by the direction from which the wind comes. Fire and heat radiation will spread rapidly downwind along the airflow. The airflow can carry high-temperature smoke and burning debris over long distances, igniting vegetation and various combustibles downwind in advance. If 0° ≤ sliding window Wind direction at start timestamp ≤22.5°, determined to be a northerly wind, the wind blows from north to south, the leeward side is the south side, the origin of the forest area. Adjacent area to due south Forming pairs; If 22.5° < sliding window Wind direction at start timestamp <67.5°, determined to be a northeasterly wind, blowing from northeast to southwest, with the southwest side being the leeward side. The origin point of the forest area is determined. Adjacent area to the southwest Forming pairs; If 67.5° ≤ sliding window Wind direction at start timestamp ≤112.5°, determined to be easterly, wind blows from east to west, downwind is the west side, the origin of the forest area. Adjacent area to due west Forming pairs; If 112.5° < sliding window Wind direction at start timestamp <157.5°, determined to be a southeast wind, blowing from southeast to northwest, with the northwest side being the downwind side. The origin point of the forest area is determined. Adjacent area to the northwest Forming pairs; If 157.5° ≤ sliding window Wind direction at start timestamp ≤202.5°, determined to be a southerly wind, the wind blows from south to north, the leeward side is the north side, the origin of the forest area is determined. Adjacent area to due north Forming pairs; If 202.5° < sliding window Wind direction at start timestamp <247.5°, determined to be a southwest wind, blowing from southwest to northeast, with the northeast side being the downwind side. The origin point of the forest area is determined. Adjacent area to the northeast Forming pairs; If 247.5° ≤ sliding window Wind direction at start timestamp ≤292.5°, determined to be a westerly wind, the wind blows from west to east, the leeward side is the east side, the origin of the forest area is set. Adjacent area to due east Forming pairs; If 292.5° < sliding window Wind direction at start timestamp <337.5°, determined to be a northwest wind, blowing from northwest to southeast, with the southeast side being the leeward side. The origin point of the forest area is determined. Adjacent area to the southeast Forming pairs; If 337.5° ≤ sliding window Wind direction at start timestamp ≤360°, determined to be a northerly wind, the wind blows from north to south, the leeward side is the south side, the origin of the forest area. Adjacent area to due south Forming pairs; Specifically, taking the wind direction angle of the starting timestamp of the sliding window as an example, the origin of the forest area can be realized by dividing the standard 8-directional wind direction angle interval within the range of 0°-360°. The precise and unique pairing with the adjacent area in the corresponding direction not only conforms to the common wind direction determination logic in the meteorological field, but also ensures the clarity and consistency of the pairing relationship. It can stably and efficiently complete the target pairing task and provide a reliable basis for subsequent prediction of the spread trend of combustibles. The following is the experimental data for the region pairing group, as shown in Table 2: Table 2: Experimental Data for the Region Pairing Group In Table 2, the experimental data for the paired-group regions were selected as a 12-minute sliding window. The experiment lasted for one minute, with data collection performed every minute. Based on the assessment, the sliding window... Relative range of thermal radiation intensity at the start timestamp This indicates that there are differences in the intensity of thermal radiation in different areas of the forest. Region 3 (78.5 W / m²), which has the highest intensity of thermal radiation in the forest, was selected as the origin point of the forest region. The wind direction angle is 67.5° < the starting timestamp < 112.5°, therefore it is determined to be an easterly wind. Region 3 (the origin of the forest region) will be included. ) and the area adjacent to due west Forming pairs; The combustible risk assessment module is used to assess the combustible development trend within the paired regional area and generate a corresponding risk index. The calculation process is as follows: S31, Given a sliding window Within this dataset, all timestamps show a southerly wind. Then, based on the environmental dataset, the origin point for the forest region is extracted. Adjacent area to due north Environmental sensing data; S32, For sliding windows The origin of the forest area The load index for each timestamp is denoted as: The adjacent area to the due north direction The load index for each timestamp is denoted as: Then calculate the sliding window. Within, the origin of the forest area Adjacent area to due north Mean of absolute difference in load index Its expression is as follows: In the formula, Indicates the origin of the forest area No. The load index of each timestamp , Indicates the adjacent area in the due north direction No. The load index of each timestamp Indicates the first The absolute difference in load index between two regions at each timestamp; S33, For sliding windows The origin of the forest area The surface temperature at each timestamp is recorded as follows: The origin of the forest area The canopy temperature at each timestamp is recorded as follows: The adjacent area to the due north direction The surface temperature at each timestamp is recorded as follows: The adjacent area to the due north direction The canopy temperature at each timestamp is recorded as follows: Then calculate the sliding window. Within, the origin of the forest area Surface temperature change rate Forest area origin Canopy temperature change rate Adjacent area to the due north direction Surface temperature change rate Adjacent area to the due north direction Canopy temperature change rate and the origin of the forest area Adjacent area to due north absolute difference in the rate of change of surface temperature absolute difference between the rate of change of canopy temperature and Its expression is as follows: S34, For sliding windows The origin of the forest area The wind speed at each timestamp is recorded as follows: The adjacent area to the due north direction The wind speed at each timestamp is recorded as follows: Then calculate the sliding window. Within, the origin of the forest area average wind speed Adjacent area to the due north direction average wind speed and the origin of the forest area Adjacent area to due north absolute difference in average wind speed Its expression is as follows: In the formula, Indicates the origin of the forest area No. Wind speed at each timestamp , Indicates the adjacent area in the due north direction No. Wind speed at each timestamp; S35. Based on S31-S34, calculate the sliding window using a weighted method. Within, the origin of the forest area Adjacent area to due north Risk index of paired groups Its expression is as follows: In the formula, The weights representing the mean of the absolute differences in the load index are... The weights representing the absolute differences in the rates of change of surface temperature. The weights representing the absolute differences in the rate of change of canopy temperature. The weights representing the absolute differences in average wind speeds , , and All are constants, and ; Specifically, by clearly calculating the absolute difference between key risk indicators in two regions and combining it with flexible weighting, a comprehensive quantification of regional risk can be achieved. This not only reflects the contribution of different risk factors but also outputs an intuitive and comparable risk index. ; The following is the experimental data for the risk index, as shown in Table 3: Table 3: Experimental Data for the Risk Index Table 3 shows the experimental data for the risk index. Region pairing group VI was selected as the experimental target. Region pairing group VI includes regions C and D. (Sliding window...) Set to 5 minutes, with data collection performed every minute; Risk Index The weights are set as follows: , , , ; The load control module is set with a fixed range of risk thresholds. This method is used to quickly determine the fire risk level of paired groups in different forest areas, providing a quantitative reference for the formulation of forest fire prevention and control strategies. The rationality of the range directly affects the accuracy of fire risk level assessment and the timeliness of prevention and control intervention. An excessively narrow risk threshold range... This could lead to overly strict fire hazard classification, misclassifying Level 1 areas with no obvious flammability as Level 2 or 3, increasing unnecessary costs for firebreaks, fire extinguishing equipment, and manpower, while also causing unnecessary panic among people in the area. Conversely, overly strict classification would blur the lines between Level 1 and Level 2, and between Level 2 and Level 3, making it impossible to distinguish the differences in flammability, thus creating a significant risk of forest fire spread. Therefore, the optimal range for this threshold interval needs to be determined through calibration experiments of the following system: Risk Threshold Interval. The calibration method is as follows: Using a historical forest fire prevention database, paired samples of forest areas with different fire risk levels were screened, covering situations with no obvious flammability trend, flammability trend, obvious flammability trend, and various terrain, vegetation types, and climatic conditions. Combustible material monitoring data (such as the thickness of fallen leaves and branches, vegetation moisture content, and the proportion of combustible material types), historical fire occurrence records, meteorological data (such as temperature, humidity, and wind speed), and data on the effectiveness of prevention and control interventions (such as fire containment success rate and spread control effectiveness) were extracted from the paired samples. Different candidate threshold ranges were set. In each calibration experiment, the fire risk level of the paired samples was classified according to the candidate threshold range. The matching degree between the classification results and the actual gold standard for fire risk assessment was recorded. Combined with dynamic risk monitoring data, the risk index of the paired areas under different prevention and control intervention intensities was simulated. The changing trend was analyzed, and the upper and lower limits of the candidate threshold intervals were adjusted. Multiple verification experiments were conducted to record the impact of threshold interval settings on the assessment of prevention and control effectiveness. For each candidate threshold interval, the collected sample data and dynamic simulation results were used as inputs to count the number of times low-level fire risk was misjudged as high-level fire risk due to improper interval settings (counted as over-assessment), the number of times high-level fire risk was misjudged as low-level fire risk (counted as under-assessment), and the degree of fit between the fire risk level classification results and the actual prevention and control effectiveness (such as fire incidence rate, early warning accuracy rate, and prevention and control resource utilization rate). Finally, the interval range that minimizes both the over-assessment rate and the under-assessment rate and has the highest degree of fit with the actual prevention and control effectiveness was selected as the risk threshold interval. The preferred range; In the experimental data of the risk index in Table 3, the risk threshold range The preferred range is defined as 0.5-0.8. Based on the judgment, the sliding window... Within, the risk index of region-paired group VI A value <0.5 indicates no significant flammability within the paired area, with a fire hazard level of 1. Response measures include continuing the current sliding window. The patrol frequency is such that there is no need to set up additional firebreaks; The ecological regulation module is set with a fixed range of thresholds to determine the load level of each forest area and the fire risk level of the paired area groups, output the assessment results and trigger corresponding ecological regulation response measures. The load level assessment process is as follows: The upper limit of the load threshold range is denoted as The lower limit of the load threshold range is denoted as ; If the first Regional load index < , indicating the first The area has low combustible material accumulation, is difficult to retain, and has weak flammability; its load level is Level 1. Response measures include maintaining the patrol frequency of the current assessment cycle, increasing the patrol frequency of the next assessment cycle, and implementing routine combustible material monitoring. No additional investment in thermal infrared imaging monitoring equipment is required. ≤th Regional load index ≤ , indicating the first The area has a moderate accumulation of combustibles, moderate retention difficulty, and moderate flammability, with a load level of 2. Response measures include increasing the frequency of patrols and combustible cleanup at the current timestamp, prioritizing manual and mechanical brush cutting and pruning to reduce surface and canopy combustible loads, and investing additional resources in thermal infrared imaging monitoring equipment for dynamic monitoring of key areas. If the [unclear text - possibly a missing section]... Regional load index > , indicating the first The area has a high accumulation of combustibles, is easy to retain, and is highly flammable, with a load level of 3. Response measures include immediately carrying out comprehensive combustible cleanup, developing a burning plan to optimize forest stand density and vegetation structure within the area, and investing additional resources in thermal infrared imaging monitoring equipment to implement daily full-coverage patrols until the [date missing]. Regional load index < ; The fire hazard assessment process is as follows: Let the upper limit of the risk threshold range be denoted as The lower limit of the risk threshold range is denoted as ; For sliding windows Within, if the origin of the forest area Adjacent area to due north Risk index of paired groups < This indicates that there is no obvious flammability trend within the area of ​​the paired group, the fire hazard level is 1, and the response measures include continuing the current sliding window. The patrol frequency is sufficient, eliminating the need for additional firebreaks. ≤ Forest region origin Adjacent area to due north Risk index of paired groups ≤ This indicates that there is a flammable tendency within the paired area, with a fire hazard level of 2. Response measures include increasing the sliding window. Increase patrol frequency within the area, activate preventative combustible material control, prepare fire extinguishing equipment, and establish additional firebreaks. (If the forest area is at its origin...) Adjacent area to due north Risk index of paired groups > This indicates a significant flammability trend within the paired area, with a fire risk level of 3. Response measures include real-time, full-coverage patrols, increasing the deployment of fire extinguishers and firebreaks, timely issuance of fire warnings, and prohibiting all open flame operations within the area, extending to the forest's origin point. Adjacent area to due north Risk index of paired groups < ; The ecological regulation module targets the entire forest area. When the load level of three adjacent areas all reaches level 2 or above, or the fire risk level all reaches level 2 or above, comprehensive ecological regulation response measures are activated. These measures include promptly notifying management personnel to create firebreaks with a width of ≥10 meters and a depth of ≥30 centimeters. The surface of the firebreaks is cleaned twice a week. Combined with planned burning and mechanical cleanup measures, the combustible load in the firebreaks and surrounding areas is reduced. The surface of the firebreaks is cleaned twice a week. After cleanup, the regional load index needs to be assessed. and risk index If the results of two consecutive assessments drop below level 2, the frequency of cleanup can be appropriately reduced; otherwise, the existing comprehensive ecological regulation and response measures should continue to be implemented.

[0020] In this embodiment, the multimodal acquisition module adopts a 20m×20m grid partitioning, deploying monitoring points according to the principle of uniform distribution. The monitoring density is flexibly adjusted for complex terrain, effectively reducing sampling errors and accurately reflecting the true distribution of combustibles. It comprehensively acquires forest topography, vegetation management, and environmental sensor data, classifies and constructs standardized datasets, and provides accurate information for assessing the ecological regulation effect of combustibles. The regional load assessment module integrates the superimposed effects of topography, vegetation, and environmental conditions to quantify combustible accumulation and flammability, outputting the corresponding load index. This provides a clear spatial basis for tiered regulation. The regional pairing module dynamically constructs pairing relationships between the origin of the fire source and its downwind neighboring area based on wind direction, accurately locating the regional origin. It also completes the pairing of adjacent areas, identifies potential fire spread paths, and the combustible risk assessment module evaluates the combustible development trend within the paired area group, outputting the corresponding risk index. This enables an upgrade from static load assessment to dynamic risk prediction. The ecological regulation module has a fixed range of thresholds to determine the load level of each forest area and the fire risk level of the paired area groups. It outputs the assessment results and triggers corresponding ecological regulation response measures. Through active intervention measures such as cleanup, burning, and isolation, it achieves long-term optimization of combustibles and closed-loop management of fire risks.

[0021] The threshold is set to facilitate comparison. The size of the threshold depends on the amount of sample data and the number of bases set by those skilled in the art for each set of sample data; as long as it does not affect the ratio between the parameter and the quantized value, it is acceptable.

[0022] The above formulas are all derived from software simulation using a large amount of data and are selected to be close to the actual values. The coefficients in the formulas are set by those skilled in the art according to the actual situation. The above description is only a preferred embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any equivalent substitutions or changes made by those skilled in the art within the technical scope disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the protection scope of the present invention.

Claims

1. A forest combustible ecological regulation effect evaluation system based on multimodal data, characterized in that, include: The multimodal acquisition module is used to acquire topographic data, vegetation management data, and environmental sensing data from all areas of the forest, and classify them into forest datasets, vegetation datasets, and environmental datasets. The regional load assessment module is used to comprehensively evaluate the combined effects of topography, vegetation, and environmental conditions in different forest areas on the accumulation, retention, and flammability of combustibles, and outputs the corresponding load index. ; The region matching module has a sliding window with a fixed duration. Generate corresponding area pairing groups according to wind direction; The combustible risk assessment module is used to assess the combustible development trend within the paired area and output the corresponding risk index. ; The ecological regulation module is set with a fixed range of thresholds to determine the load level of each forest area and the fire risk level of the paired regional groups, output the assessment results and trigger corresponding ecological regulation response measures.

2. The forest combustible ecological regulation effect evaluation system based on multimodal data according to claim 1, characterized in that: Each area of ​​the forest has several monitoring points, and the forest dataset includes the total number of monitoring points, altitude, slope, and aspect angle for each area.

3. The forest combustible ecological regulation effect evaluation system based on multimodal data according to claim 2, characterized in that: The vegetation dataset includes aboveground biomass, dead combustible load, humus layer thickness, humus layer moisture content, moss and lichen layer thickness, moss and lichen layer moisture content, herbaceous layer cover, live combustible load, shrub layer cover, shrub layer dead branches, tree layer canopy closure, and tree layer dead branches for each region.

4. The forest combustible ecological regulation effect evaluation system based on multimodal data according to claim 3, characterized in that: The environmental dataset includes the average daily sunshine duration, thermal radiation intensity, surface temperature, canopy temperature, and wind speed for each region.

5. The forest combustible ecological regulation effect evaluation system based on multimodal data according to claim 4, characterized in that: The load index The calculation process is as follows: S11. Based on the forest dataset, vegetation dataset, and environment dataset, extract the first... Topographic data, vegetation management data, and environmental sensor data of the region at the current timestamp; S12, Calculate the first Average altitude of the region at the current timestamp and standard deviation ; S13, Calculate the... Average slope of the current timestamp in the region and standard deviation ; S14. Calculate the weighted average. Reserve index of the region at the current timestamp ; S15, Determine the first The slope aspect type of each monitoring point in the area was determined, and the first... The proportion of shady slopes in the area The proportion of sunny slopes and the proportion of semi-sunny slopes ; S16. Calculate the weighted average. Decomposition adjustment index of the current timestamp of the region ; S17, Calculate the... Average thickness of humus layer at the current timestamp of the region Average moisture content of humus layer Mean thickness of moss and lichen layer Mean moisture content of moss and lichen layers and surface humus index ; S18. Calculate the first using a weighted method. Distribution index of current timestamp in the region ; S19. Based on S11-S18, calculate the first... Load index of the region at the current timestamp .

6. The forest combustible ecological regulation effect evaluation system based on multimodal data according to claim 5, characterized in that: The region pairing process is as follows: S21. Extract the sliding window based on the environment dataset. Within, environmental sensor data for each area of ​​the forest; S22, Calculate the sliding window Relative range of thermal radiation intensity at the start timestamp ; S23, Based on the sliding window Relative range of thermal radiation intensity at the start timestamp Determine the origin of the forest area The process is as follows: If sliding window Relative range of thermal radiation intensity at the start timestamp This indicates that the thermal radiation intensity is completely equal across all areas of the forest, and the geometric center of the forest area is selected as the origin of the forest region. ; If sliding window Relative range of thermal radiation intensity at the start timestamp This indicates that the intensity of thermal radiation varies across different areas of the forest; the area with the highest intensity of thermal radiation is selected as the origin point of the forest area. If multiple regions with the highest thermal radiation intensity exist, select the region with the shortest straight-line distance to the geometric center of the forest area as the origin of the forest region. ; S24. Determine the wind direction type for each timestamp and establish the origin for the forest area. The pairing process is as follows: The origin of the forest area The adjacent area to due north is denoted as The origin of the forest area The adjacent area to the northeast is denoted as The origin of the forest area The adjacent area to the due east direction is denoted as The origin of the forest area The adjacent area to the southeast is denoted as The origin of the forest area The adjacent area to the due south direction is denoted as The origin of the forest area The adjacent area to the southwest is denoted as The origin of the forest area The adjacent area to the due west direction is denoted as The origin of the forest area The adjacent area to the northwest is denoted as ; If 0° ≤ sliding window Wind direction at start timestamp ≤22.5°, determined to be a northerly wind, the origin point of the forest area is set. Adjacent area to due south Forming pairs; If 22.5° < sliding window Wind direction at start timestamp <67.5°, determined to be a northeasterly wind, the origin point of the forest area. Adjacent area to the southwest Forming pairs; If 67.5° ≤ sliding window Wind direction at start timestamp ≤112.5°, determined to be easterly, the origin point of the forest area is set. Adjacent area to due west Forming pairs; If 112.5° < sliding window Wind direction at start timestamp <157.5°, determined to be a southeast wind, the origin point of the forest area is set. Adjacent area to the northwest Forming pairs; If 157.5° ≤ sliding window Wind direction at start timestamp ≤202.5°, determined to be a southerly wind, the origin point of the forest area is set. Adjacent area to due north Forming pairs; If 202.5° < sliding window Wind direction at start timestamp <247.5°, determined to be a southwest wind, the origin point of the forest area is set. Adjacent area to the northeast Forming pairs; If 247.5° ≤ sliding window Wind direction at start timestamp ≤292.5°, determined to be westerly wind, the origin point of the forest area is set. Adjacent area to due east Forming pairs; If 292.5° < sliding window Wind direction at start timestamp <337.5°, determined to be a northwest wind, the origin point of the forest area is set. Adjacent area to the southeast Forming pairs; If 337.5° ≤ sliding window Wind direction at start timestamp ≤360°, determined to be a northerly wind, the origin of the forest area is... Adjacent area to due south Form pairs.

7. The forest combustible ecological regulation effect evaluation system based on multimodal data according to claim 6, characterized in that: The risk index The calculation process is as follows: S31, Given a sliding window Within this dataset, all timestamps show a southerly wind. Then, based on the environmental dataset, the origin point for the forest region is extracted. Adjacent area to due north Environmental sensing data; S32, Calculate the sliding window Within, the origin of the forest area Adjacent area to due north Mean of absolute difference in load index ; S33, Calculate the sliding window Within, the origin of the forest area Surface temperature change rate Forest area origin Canopy temperature change rate Adjacent area to the due north direction Surface temperature change rate Adjacent area to the due north direction Canopy temperature change rate and the origin of the forest area Adjacent area to due north absolute difference in the rate of change of surface temperature absolute difference between the rate of change of canopy temperature and ; S34, Calculate the sliding window Within, the origin of the forest area average wind speed Adjacent area to the due north direction average wind speed and the origin of the forest area Adjacent area to due north absolute difference in average wind speed ; S35. Based on S31-S34, calculate the sliding window using a weighted method. Within, the origin of the forest area Adjacent area to due north Risk index of paired groups .

8. The forest combustible ecological regulation effect evaluation system based on multimodal data according to claim 7, characterized in that: The load level assessment process is as follows: The upper limit of the load threshold range is denoted as The lower limit of the load threshold range is denoted as ; If the first Regional load index < , indicating the first The area has low combustible material accumulation, is difficult to retain, and has weak flammability; its load level is Level 1. Response measures include maintaining the patrol frequency of the current assessment cycle, increasing the patrol frequency of the next assessment cycle, and implementing routine combustible material monitoring. No additional investment in thermal infrared imaging monitoring equipment is required. ≤th Regional load index ≤ , indicating the first The area has a moderate accumulation of combustibles, moderate retention difficulty, and moderate flammability, with a load level of 2. Response measures include increasing the frequency of patrols and combustible cleanup at the current timestamp, prioritizing manual and mechanical brush cutting and pruning to reduce surface and canopy combustible loads, and investing additional resources in thermal infrared imaging monitoring equipment for dynamic monitoring of key areas. If the [unclear text - possibly a missing section]... Regional load index > , indicating the first The area has a high accumulation of combustibles, is easy to retain, and is highly flammable, with a load level of 3. Response measures include immediately carrying out comprehensive combustible cleanup, developing a burning plan to optimize forest stand density and vegetation structure within the area, and investing additional resources in thermal infrared imaging monitoring equipment to implement daily full-coverage patrols until the [date missing]. Regional load index < .

9. The forest combustible ecological regulation effect evaluation system based on multimodal data according to claim 8, characterized in that: The fire hazard assessment process is as follows: Let the upper limit of the risk threshold range be denoted as The lower limit of the risk threshold range is denoted as ; For sliding windows Within, if the origin of the forest area Adjacent area to due north Risk index of paired groups < This indicates that there is no obvious flammability trend within the area of ​​the paired group, the fire hazard level is 1, and the response measures include continuing the current sliding window. The patrol frequency is sufficient, eliminating the need for additional firebreaks. ≤ Forest region origin Adjacent area to due north Risk index of paired groups ≤ This indicates that there is a flammable tendency within the paired area, with a fire hazard level of 2. Response measures include increasing the sliding window. Increase patrol frequency within the area, activate preventative combustible material control, prepare fire extinguishing equipment, and establish additional firebreaks. (If the forest area is at its origin...) Adjacent area to due north Risk index of paired groups > This indicates a significant flammability trend within the paired area, with a fire risk level of 3. Response measures include real-time, full-coverage patrols, increasing the deployment of fire extinguishers and firebreaks, timely issuance of fire warnings, and prohibiting all open flame operations within the area, extending to the forest's origin point. Adjacent area to due north Risk index of paired groups < .

10. The forest combustible ecological regulation effect evaluation system based on multimodal data according to claim 9, characterized in that: The ecological regulation module targets the entire forest area. When the load level of three adjacent areas all reaches level 2 or above, or the fire risk level all reaches level 2 or above, comprehensive ecological regulation response measures are activated. These measures include promptly notifying management personnel to create firebreaks with a width of ≥10 meters and a depth of ≥30 centimeters, and combining planned burning and mechanical clearing measures to reduce the combustible load in the firebreaks and surrounding areas. The surface of the firebreaks is cleared twice a week, and the regional load index is assessed after clearing. and risk index If the results of two consecutive assessments drop below level 2, the frequency of cleanup can be appropriately reduced; otherwise, the existing comprehensive ecological regulation and response measures should continue to be implemented.

Images