Forest fire risk assessment method and system based on big data analysis

Abstract

The invention provides a forest fire risk assessment method and system based on big data analysis, and the method comprises the steps: collecting data from a plurality of channels, carrying out the cleaning and formatting conversion of the original data, so as to guarantee the accuracy in a subsequent analysis process, correspondingly planning a proportion map of a monitoring region and a historical risk region according to historical records, and carrying out the calculation of the proportion map. The method comprises the following steps: constructing corresponding characteristic variables according to known key factors influencing the occurrence of forest fire, carrying out conversion calculation on the characteristic variables, extracting corresponding characteristic factors according to historical accident records after the calculation is completed, and synchronously amplifying and reducing the characteristic factors according to corresponding risk proportions, so as to obtain the risk of forest fire. And participating in subsequent prediction calculation as a specific variable. By comprehensively analyzing multi-source data, the system can more accurately identify a high-risk area and dynamically adjust the monitoring proportion, the smoothness and prediction precision of the data are further improved, prevention measures can be taken in time, and fire loss is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of forest fire early warning, and in particular to a forest fire risk assessment method and system based on big data analysis. Background Art

[0002] With the impact of global climate change and human activities, the frequency and scale of forest fires have shown an increasing trend year by year. Traditional forest fire management methods mainly rely on manual inspections and empirical judgments, which are not only inefficient but also difficult to cope with large-scale and sudden fire events. In recent years, with the development of technologies such as big data, the Internet of Things, and artificial intelligence, forest fire risk assessment methods based on big data analysis have gradually become a research hotspot. Modern forest fire risk assessment systems also integrate advanced GIS technology and visualization tools, which can intuitively display risk maps to help decision makers quickly understand the current situation and take corresponding actions. However, due to the many causes of fire, it is difficult to match the corresponding states during multi-source data monitoring, resulting in reduced accuracy and flexibility.

[0003] Therefore, in response to the above problems, a forest fire risk assessment method and system based on big data analysis is now developed. Summary of the invention

[0004] In order to overcome the shortcomings of the prior art that due to the large number of fire causes, it is difficult to match the corresponding states during multi-source data monitoring, resulting in reduced accuracy and flexibility, the present invention provides a forest fire risk assessment method and system based on big data analysis.

[0005] The technical solution of the present invention is: a forest fire risk assessment method based on big data analysis, comprising: Step 1: Collect data from multiple channels, including but not limited to satellite remote sensing images, ground sensor monitoring data, temperature, humidity, wind speed meteorological parameters provided by weather stations, vegetation type distribution maps, topographic features, and historical fire event records of the areas to be monitored in the past few decades. Clean and format these raw data to ensure accuracy in the subsequent analysis process, and then plan a map of the proportion of monitoring areas and historical risk areas based on historical records; Step 2: Based on the known key factors that affect the occurrence of forest fires, construct corresponding characteristic variables and perform transformation calculations on these characteristic variables. After the calculations are completed, extract the corresponding characteristic factors based on historical accident records, and simultaneously amplify and reduce these characteristic factors according to the corresponding risk ratios to obtain standardized numerical parameters, which are used as variables in subsequent prediction calculations; Step 3: The obtained variables are quadratically standardized according to the parameter distribution ratio in each monitoring area to obtain the final impact data after the corresponding proportion. Through learning and calculation, the proportion of each monitoring area is dynamically adjusted, and the key states and areas are planned. This state and area are also dynamically controlled according to the final impact data, and real-time planning and adjustment are carried out; Step 4: Calculate the risk probability of different areas according to the predicted fire trend, draw the fire line inflection point and control curve, set up at least five control curves according to the forest area, and mark the points on the topographic map. After the marking is completed, the operator will confirm and verify the marked points on site and explore the feasibility of installing the corresponding equipment; Step 5: Update input data regularly to keep the model up to date. At the same time, as new fire incidents occur, continuously adjust and improve the existing model to improve prediction accuracy.

[0006] As a preferred embodiment of the present invention, in the process of calculating the characteristic variables, the key factors are first defined, including but not limited to temperature T, humidity H, wind speed W, normalized difference vegetation index V, slope S, and slope aspect A. The minimum-maximum normalization method is used to define these data to ensure that the data are on the same scale and determine the corresponding historical extreme value range.

[0007] As a preferred embodiment of the present invention, the standardized characteristic variables are comprehensively composed, and a weight vector is defined according to the influence of the characteristic variables under different regional conditions. The weight vector needs to be adjusted according to the corresponding regional environment ratio, and the adjustment interval is within 0-2. When the comprehensive calculation is performed, the weight vector is brought in for calculation and expression: ,in, , , , , and The data after the definitions of temperature T, humidity H, wind speed W, normalized difference vegetation index V, slope S, and slope aspect A are respectively expressed. , , , , and are the weight vectors corresponding to temperature T, humidity H, wind speed W, normalized difference vegetation index V, slope S, and aspect A respectively; After obtaining the preliminary comprehensive feature vector, the time series is introduced for analysis, and the exponential weighted moving average method is used to smooth the data. The formula is continuously updated and the smoothing factor is introduced. , where 0<α<1.

[0008] As a preferred embodiment of the present invention, the formula for calculating the fire occurrence probability coefficient is: ,in and It is the parameter obtained by fitting and adjusting according to the changes in the time series. e is the base of the natural logarithm, and F is the comprehensive eigenvector.

[0009] As a preferred embodiment of the present invention, the formula for calculating the fire trend is: ; in represents the flame area density, Indicates the change of fire intensity over time. represents the velocity vector of the flame front, It indicates the changing trend value of flame in spatial state, so as to assist the system to obtain the state of fire direction and diffusion range.

[0010] As a preferred embodiment of the present invention, the method of dynamically adjusting the scale of each monitoring area is as follows: first, the entire study area is divided into multiple monitoring areas, and each grid unit is represented as , where i and j are the row and column indices of the grid, setting an initial scale for each monitoring area , define the variables of each monitoring area as , then the value of the Kth position is , the parameter distribution ratio in each monitoring area is , then the formula for the secondary norm is: ,in is the characteristic variable after quadratic normalization, m is the number of monitoring areas; Secondly, a loss function L is defined to measure the difference between the current prediction and the actual fire situation. The loss function formula is: ,in is the percentage of actual fire occurrences in historical data. is the probability data predicted by the model, that is, according to the formula The probability value obtained; Next, the gradient descent method is used in the update process, and the specific expression formula is: , in is the loss gradient, The probability predicted by the model The partial derivative of the ratio value at position t with an initial ratio set for each monitoring area indicates how the probability predicted by the model changes with the change of the ratio value; Finally, key states and areas are determined based on the final impact data and the adjusted ratios, and real-time planning and adjustments are carried out to ensure that resources are allocated to high-risk areas first.

[0011] A forest fire risk assessment system based on big data analysis, including: A data collection layer, which is responsible for acquiring required data from various sources, including remote sensing satellites, ground observation stations, and public databases; The data processing layer performs data cleaning, integration, and conversion tasks to prepare clean and reliable data for the next step of analysis; An analysis engine layer, which includes a feature extraction module, a model training module, and a risk assessment module; The user interface layer is used to provide users with intuitive map displays and risk reporting functions to help decision makers quickly understand the current situation and take corresponding actions.

[0012] As a preferred embodiment of the present invention, it also includes a detection device and a reaction device, wherein the detection device is used to assist the system in risk monitoring and early warning, and the reaction device is used to assist in fire extinguishing in an emergency.

[0013] By adopting the above technical solution, the present invention has the following advantages: 1. Through comprehensive analysis of multi-source data, the system can more accurately identify high-risk areas and dynamically adjust the monitoring ratio. It uses time series analysis methods such as exponentially weighted moving average to further improve the smoothness of data and prediction accuracy. This multi-level data fusion and dynamic adjustment mechanism makes forest fire risk assessment more accurate, helps to take preventive measures in a timely manner, and reduces fire losses.

[0014] 2. The present invention can not only provide real-time risk assessment results, but also calculate the prevention and control curve according to the direction of the fire, and form point marks on the topographic map. After these marked points are confirmed on site, they can guide operators to carry out effective fire protection device installation and resource allocation.

[0015] 3. The present invention includes detection equipment and reaction equipment, which can respond quickly in emergency situations and assist in fire fighting, thereby greatly improving the efficiency of emergency handling, ensuring that resources can be allocated to the places where they are most needed, and achieving optimal allocation of resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a flowchart of the steps of the forest fire risk assessment method based on big data analysis of the present invention.

[0017] Figure 2 It is a structural schematic diagram of the forest fire risk assessment system based on big data analysis of the present invention.

[0018] Figure 3 A flow chart of the steps of drawing the fire line inflection point and the prevention and control curve is proposed for the present invention.

[0019] Figure 4 A schematic diagram of the fire line inflection point and the prevention and control curve is drawn for the present invention. DETAILED DESCRIPTION

[0020] Reference to an embodiment herein means that a particular feature, structure, or characteristic described in conjunction with the embodiment may be included in at least one embodiment of the present invention. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0021] A forest fire risk assessment method based on big data analysis, such as Figure 1 As shown, including: Step 1: Collect data from multiple channels, including but not limited to satellite remote sensing images, ground sensor monitoring data, temperature, humidity, wind speed meteorological parameters provided by weather stations, vegetation type distribution maps, topographic features, and historical fire event records of the areas to be monitored in the past few decades. Clean and format these raw data to ensure accuracy in the subsequent analysis process, and then plan a map of the proportion of monitoring areas and historical risk areas based on historical records; Step 2: Based on the known key factors that affect the occurrence of forest fires, construct corresponding characteristic variables and perform transformation calculations on these characteristic variables. After the calculations are completed, extract the corresponding characteristic factors based on historical accident records, and simultaneously amplify and reduce these characteristic factors according to the corresponding risk ratios to obtain standardized numerical parameters, which are used as variables in subsequent prediction calculations; Step 3: The obtained variables are quadratically standardized according to the parameter distribution ratio in each monitoring area to obtain the final impact data after the corresponding proportion. Through learning and calculation, the proportion of each monitoring area is dynamically adjusted, and the key states and areas are planned. This state and area are also dynamically controlled according to the final impact data, and real-time planning and adjustment are carried out; Step 4: Calculate the risk probability of different areas according to the predicted fire trend, draw the fire line inflection point and control curve, set up at least five control curves according to the forest area, and mark the points on the topographic map. After the marking is completed, the operator will confirm and verify the marked points on site and explore the feasibility of installing the corresponding equipment; Step 5: Update input data regularly to keep the model up to date. At the same time, as new fire incidents occur, continuously adjust and improve the existing model to improve prediction accuracy.

[0022] It should be noted that when drawing the fire line inflection point and the prevention and control curve, Figure 3 As shown, the method used is as follows: Collect data through satellite imagery, drone surveillance, ground observations, and use GIS analysis tools to identify and mark fire front inflection points; After the marking of the turning point is determined, a comprehensive analysis is designed based on factors such as the forest farm's topography, vegetation type, wind direction, etc., while taking into account the effective deployment of emergency response resources, and combining historical fire data and simulation prediction models to optimize the location of the prevention and control curve; When setting up control curves, the first control curve should be as close as possible to the known or potential fire source, and the subsequent control curves should gradually move away from the fire source to form a gradient protection system. The distance between each control curve needs to be adjusted according to the specific size of the forest farm and the results of risk assessment; Finally, the system directly generates the corresponding line map, such as Figure 4 As shown, relevant personnel rely on the line chart to set up the corresponding prevention and control curve.

[0023] In the process of calculating the characteristic variables, the key factors are first defined, including but not limited to temperature T, humidity H, wind speed W, normalized difference vegetation index V, slope S, and slope aspect A. The minimum-maximum normalization method is used to define these data to ensure that the data are on the same scale and determine the corresponding historical extreme value range.

[0024] The standardized characteristic variables are synthesized, and the weight vector is defined according to the influence of the characteristic variables under different regional conditions. The weight vector needs to be adjusted according to the corresponding regional environment proportion, and the adjustment interval is within 0-2. When calculating comprehensively, the weight vector is brought in for calculation: ,in, , , , , and The data after the definitions of temperature T, humidity H, wind speed W, normalized difference vegetation index V, slope S, and slope aspect A are respectively expressed. , , , , and are the weight vectors corresponding to temperature T, humidity H, wind speed W, normalized difference vegetation index V, slope S, and aspect A respectively; After obtaining the preliminary comprehensive feature vector, the time series is introduced for analysis, and the exponential weighted moving average method is used to smooth the data. The formula is continuously updated and the smoothing factor is introduced. , where 0<α<1.

[0025] The formula for calculating the fire occurrence probability coefficient is: ,in and It is the parameter obtained by fitting and adjusting according to the changes in the time series. e is the base of the natural logarithm, and F is the comprehensive eigenvector.

[0026] The formula for calculating the direction of fire is: ; in represents the flame area density, Indicates the change of fire intensity over time. represents the velocity vector of the flame front, It indicates the changing trend value of flame in spatial state, so as to assist the system to obtain the state of fire direction and diffusion range.

[0027] The way to dynamically adjust the scale of each monitoring area is as follows: first, the entire study area is divided into multiple monitoring areas, and each grid unit is represented as , where i and j are the row and column indices of the grid, setting an initial scale for each monitoring area , define the variables of each monitoring area as , then the value of the Kth position is , the parameter distribution ratio in each monitoring area is , then the formula for the quadratic norm is ,in is the characteristic variable after quadratic normalization, m is the number of monitoring areas; Secondly, a loss function L is defined to measure the difference between the current prediction and the actual fire situation. The loss function formula is: ,in is the percentage of actual fire occurrences in historical data. is the probability predicted by the model, that is, according to the formula The probability value obtained; Next, the gradient descent method is used in the update process, and the specific expression formula is: ,in is the loss gradient, The probability predicted by the model The partial derivative of the ratio value at position t with an initial ratio set for each monitoring area indicates how the probability predicted by the model changes with the change of the ratio value; Finally, key states and areas are determined based on the final impact data and the adjusted ratios, and real-time planning and adjustments are carried out to ensure that resources are allocated to high-risk areas first.

[0028] A forest fire risk assessment system based on big data analysis, such as Figure 2 As shown, including: A data collection layer, which is responsible for acquiring required data from various sources, including remote sensing satellites, ground observation stations, public databases, etc.; The data processing layer performs tasks such as data cleaning, integration, and conversion to prepare clean and reliable data for the next step of analysis; An analysis engine layer, which includes a feature extraction module, a model training module, and a risk assessment module; The user interface layer is used to provide users with intuitive map displays, risk reports and other functions to help decision makers quickly understand the current situation and take corresponding actions.

[0029] It also includes detection equipment and reaction equipment. The detection equipment is used to assist the system in risk monitoring and early warning, and the reaction equipment is used to assist in fire extinguishing in emergency situations.

[0030] The above embodiments are provided for persons familiar with the art to implement or use the present invention. Personnel familiar with the art can make various modifications or changes to the above embodiments without departing from the concept of the present invention. Therefore, the protection scope of the present invention is not limited to the above embodiments, but should be the maximum scope of the innovative features mentioned in the claims.

Claims

1. A forest fire risk assessment method based on big data analysis, characterized in that: Included are: Step 1: Collect data from multiple channels, including but not limited to satellite remote sensing images, ground sensor monitoring data, temperature, humidity, wind speed meteorological parameters provided by weather stations, vegetation type distribution maps, topographic features, and historical fire event records of the areas to be monitored in the past few decades. Clean and format these raw data to ensure accuracy in the subsequent analysis process, and then plan a map of the proportion of monitoring areas and historical risk areas based on historical records; Step 2: Based on the known key factors that affect the occurrence of forest fires, construct corresponding characteristic variables and perform transformation calculations on these characteristic variables. After the calculations are completed, extract the corresponding characteristic factors based on historical accident records, and simultaneously amplify and reduce these characteristic factors according to the corresponding risk ratios to obtain standardized numerical parameters, which are used as variables in subsequent prediction calculations; Step 3: The obtained variables are quadratically standardized according to the parameter distribution ratio in each monitoring area to obtain the final impact data after the corresponding proportion. Through learning and calculation, the proportion of each monitoring area is dynamically adjusted, and the key states and areas are planned. This state and area are also dynamically controlled according to the final impact data, and real-time planning and adjustment are carried out; Step 4: Calculate the risk probability of different areas according to the predicted fire trend, draw the fire line inflection point and control curve, set up at least five control curves according to the forest area, and mark the points on the topographic map. After the marking is completed, the operator will confirm and verify the marked points on site and explore the feasibility of installing the corresponding equipment; Step 5: Update input data regularly to keep the model up to date. At the same time, as new fire incidents occur, continuously adjust and improve the existing model to improve prediction accuracy.

2. A forest fire risk assessment method based on big data analysis according to claim 1, characterized in that: In the process of calculating the characteristic variables, the key factors are first defined, including but not limited to temperature T, humidity H, wind speed W, normalized difference vegetation index V, slope S, and slope aspect A. The minimum-maximum normalization method is used to define these data to ensure that the data are on the same scale and determine the corresponding historical extreme value range.

3. A forest fire risk assessment method based on big data analysis according to claim 1, characterized in that: The standardized characteristic variables are synthesized, and the weight vector is defined according to the influence of the characteristic variables under different regional conditions. The weight vector needs to be adjusted according to the corresponding regional environment proportion, and the adjustment interval is within 0-2. When calculating comprehensively, the weight vector is brought in for calculation: ,in, , , , , and The data after the definitions of temperature T, humidity H, wind speed W, normalized difference vegetation index V, slope S, and slope aspect A are respectively expressed. , , , , and are the weight vectors corresponding to temperature T, humidity H, wind speed W, normalized difference vegetation index V, slope S, and slope aspect A respectively; After obtaining the preliminary comprehensive feature vector, the time series is introduced for analysis, and the exponential weighted moving average method is used to smooth the data. The formula is continuously updated and the smoothing factor is introduced. , where 0 < α < 1.

4. A forest fire risk assessment method based on big data analysis according to claim 1, characterized in that: The formula for calculating the fire occurrence probability coefficient is: ,in and It is the parameter obtained by fitting and adjusting according to the changes in the time series. e is the base of the natural logarithm, and F is the comprehensive eigenvector.

5. A forest fire risk assessment method based on big data analysis according to claim 1, characterized in that: The formula for calculating the direction of fire is: ; in represents the flame area density, Indicates the change of fire intensity over time. represents the velocity vector of the flame front, It indicates the changing trend value of flame in spatial state, so as to assist the system to obtain the state of fire direction and diffusion range.

6. A forest fire risk assessment method based on big data analysis according to claim 1, characterized in that: The way to dynamically adjust the scale of each monitoring area is as follows: first, the entire study area is divided into multiple monitoring areas, and each grid unit is represented as , where i and j are the row and column indices of the grid, setting an initial scale for each monitoring area , define the variables of each monitoring area as , then the value of the Kth position is , the parameter distribution ratio in each monitoring area is , then the formula for the secondary norm is: ,in is the characteristic variable after quadratic normalization, m is the number of monitoring areas; Secondly, a loss function L is defined to measure the difference between the current prediction and the actual fire situation. The loss function formula is: ,in is the percentage of actual fire occurrences in historical data. is the probability data predicted by the model, that is, according to the formula The probability value obtained; Next, the gradient descent method is used in the update process, and the specific expression formula is: , in is the loss gradient, The probability predicted by the model The partial derivative of the ratio value at position t with an initial ratio set for each monitoring area indicates how the probability predicted by the model changes with the change of the ratio value; Finally, key states and areas are determined based on the final impact data and the adjusted ratios, and real-time planning and adjustments are carried out to ensure that resources are allocated to high-risk areas first.

7. A forest fire risk assessment system based on big data analysis, characterized in that: Included are: A data collection layer, which is responsible for acquiring required data from various sources, including remote sensing satellites, ground observation stations, and public databases; The data processing layer performs data cleaning, integration, and conversion tasks to prepare clean and reliable data for the next step of analysis; An analysis engine layer, which includes a feature extraction module, a model training module, and a risk assessment module; The user interface layer is used to provide users with intuitive map displays and risk reporting functions to help decision makers quickly understand the current situation and take corresponding actions.

8. A forest fire risk assessment system based on big data analysis according to claim 7, characterized in that: It also includes detection equipment and reaction equipment. The detection equipment is used to assist the system in risk monitoring and early warning, and the reaction equipment is used to assist in fire extinguishing in an emergency.

Images