Mountain fire evolution risk grading method and system under multi-dimensional data control

Abstract

The invention relates to the technical field of mountain fire risk grading, in particular to a mountain fire evolution risk grading method and system under multi-dimensional data control, and the method comprises the steps: carrying out the grid division of a preset region, and obtaining various fire danger factors of each grid in real time; forward processing is carried out on the fire danger factors generating negative influence on the forest fire risk, forest fire spreading characteristic values and correction coefficients of all the grids are obtained, and the forest fire spreading speed, obtained under the forest fire spreading model, of all the grids is corrected; obtaining the probability of triggering the forest fire by each grid and the conditional probability of triggering the forest fire by each fire factor value in each grid, and obtaining the characteristic factor of each grid; the stability weight of each fire danger factor of each grid is obtained, the risk score of each grid is obtained through the characteristic factor, the stability weight and the conditional probability, and the risk score is used for grading the forest fire evolution risk of each grid. The invention aims to improve the accuracy of mountain fire evolution risk grading.

Description

Technical Field

[0001] This application relates to the technical field of wildfire risk grading, and specifically to a wildfire evolution risk grading method and system under multi-dimensional data control. Background Art

[0002] A transmission line is affected by factors such as a continuous natural environment and material aging, and has a relatively high risk of triggering a wildfire. A fire caused by a power line often causes large-scale safety risks and ecological damage. Currently, in the industry, a variety of intelligent sensors are usually used to monitor various micro-environment indicators, and based on the multi-dimensional data obtained in real time, a system architecture for wildfire risk grading is established to determine important sections for wildfire prevention and control, and guide staff to conduct subsequent inspections and protection.

[0003] The current wildfire risk grading system usually evaluates the probability of a wildfire occurring based on environmental factors, ignoring the dynamic influence of various factors during the evolution process (such as the spreading stage) after a wildfire occurs, resulting in insufficient reliability of wildfire risk grading and being difficult to accurately reflect the wildfire risks existing in each region. The wildfire spreading speed is one of the important factors reflecting the degree of wildfire risk. How to introduce the wildfire spreading mechanism into the wildfire risk grading system and improve the grading performance of the wildfire risk grading system for wildfire evolution risks is an urgent problem to be solved currently. Summary of the Invention

[0004] In view of the above, it is necessary to provide a wildfire evolution risk grading method and system under multi-dimensional data control, which improves the accuracy of wildfire evolution risk grading compared with traditional wildfire evolution risk grading methods and systems: In a first aspect, an embodiment of this application provides a wildfire evolution risk grading method under multi-dimensional data control. The method includes the following steps: Divide a preset area into grids, and obtain various fire risk factors of each grid in real time, where there is a wind direction vector; For the fire risk factors that have a negative impact on wildfire risk, perform positive processing. By analyzing the dispersion of various fire risk factors of all grids, obtain the weights of various fire risk factors, and combine the various fire risk factors of each grid to obtain the wildfire spreading characteristic value of each grid; Preset each neighboring grid of each grid, obtain the position vector and distance of each neighboring grid relative to each grid, and through the similarity between the wind direction vector of each grid and the position vector, as well as the wildfire spreading characteristic value, and combine the distance, obtain the correction coefficient of each grid, and correct the wildfire spreading speed of each grid obtained under the forest fire spreading model; Using a pre-trained Bayesian classifier, obtain the probability of wildfires caused by each grid and the conditional probability of wildfires caused by the values of each fire risk factor in each grid. Based on the correction results, the probabilities, and the distances of each neighboring grid of each grid, obtain the characteristic factors of each grid; Using a comprehensive evaluation method, obtain the stable weights of each fire risk factor for each grid. Based on the characteristic factors, the stable weights, and the conditional probabilities, obtain the risk scores for each grid, where the risk scores are used to grade the wildfire evolution risks of each grid.

[0005] In one embodiment, the wildfire spread eigenvalue is the sum of the products of all fire risk factors except the wind direction vector in each grid and their weights, where the weights of each fire risk factor are obtained by an objective weighting method.

[0006] In one embodiment, the process of obtaining the correction coefficient is as follows: Calculate the mean of the wildfire spread eigenvalues of each grid and its neighboring grids; Calculate the product of the absolute value of the similarity and the mean; The correction coefficient is directly proportional to the product corresponding to each neighboring grid of each grid and inversely proportional to the distance.

[0007] In one embodiment, the expression of the correction coefficient is: ; where represents the correction coefficient of the j-th grid; N represents the number of neighboring grids of the j-th grid; represents the product corresponding to the b-th neighboring grid of the j-th grid; ε represents a preset positive number; norm( ) represents a normalization function; represents the distance between the j-th grid and its b-th neighboring grid.

[0008] In one embodiment, the correction of the wildfire spread speed of each grid obtained under the wildfire spread model includes: Take the product of the wildfire spread speed of each grid obtained under the wildfire spread model and the correction coefficient as the corrected wildfire spread speed of each grid.

[0009] In one embodiment, the expression of the characteristic factor is: ; where represents the characteristic factor of the j-th grid; norm( ) represents a normalization function; N represents the number of neighboring grids of the j-th grid; represents the corrected wildfire spread speed of the b-th neighboring grid of the j-th grid; represents the probability of the b-th neighboring grid of the j-th grid causing a wildfire; Denotes the distance between the j-th grid and its b-th neighboring grid.

[0010] In one embodiment, the process of obtaining the risk score is as follows: Calculate the product value of the stability weight and the conditional probability; Calculate the average value of the product values of all fire risk factors except the wind direction vector for each grid; Combine the average value and the characteristic factor to obtain the risk score for each grid.

[0011] In one embodiment, the risk score is the product of the average value and the characteristic factor.

[0012] In one embodiment, the method for classifying the wildfire evolution risk of each grid is as follows: Preset the value range of each wildfire evolution risk level. When the risk score of any grid is within the value range of any wildfire evolution risk level, then classify the risk of the any grid as the any wildfire evolution risk level.

[0013] In a second aspect, the embodiment of the present application also provides a wildfire evolution risk classification system under multi-dimensional data control, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of the wildfire evolution risk classification method described in any one of the above.

[0014] The present application has at least the following beneficial effects: The present application uses a variety of intelligent sensors to monitor various micro-environment indicators, and positively processes the fire risk factors collected by the intelligent sensors that have a negative impact on the wildfire risk, improving the availability of data and the efficiency and quality of monitoring and analysis; Furthermore, by analyzing the dispersion degree of each fire risk factor to obtain the weight, and then obtaining the wildfire spread characteristic value, and at the same time considering the influence of the wind direction on the wildfire spread, obtaining the correction coefficient, and correcting the wildfire spread model, the calculation accuracy of the wildfire spread speed is significantly improved under the superposition effect of a large time, space scale and multiple fire risk factors, and then the reliability of the wildfire evolution risk assessment is improved; Furthermore, considering the wildfire risk levels of each grid in both aspects of wildfire spread and wildfire occurrence, a comprehensive evaluation method is used to analyze the interaction between different fire risk factors on the impact of triggering the wildfire risk, which can more comprehensively reflect the comprehensive wildfire risk of each grid and improve the accuracy of classifying the wildfire evolution risk. Description of the Drawings

[0015] To more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0016] Figure 1 It is a flowchart of the steps of a wildfire evolution risk grading method under multi-dimensional data control provided by an embodiment of the present application; Figure 2 It is a schematic diagram of the acquisition process of risk scores. Detailed implementation manners

[0017] In the description of the embodiments of the present application, words such as "exemplary", "or", "for example" are used to represent examples, illustrations or explanations. Any embodiment or design solution described as "exemplary" or "for example" in the embodiments of the present application should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Exactly, using words such as "exemplary", "or", "for example" aims to present relevant concepts in a specific manner.

[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present application belongs. The terms used in the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. It should be understood that unless otherwise stated in the present application, " / " means "or".

[0019] In addition, it should be noted that the terms "first" and "second" in the present application are used to distinguish similar objects and are not used to describe a specific order or sequence.

[0020] The following will specifically describe the specific solutions of the wildfire evolution risk grading method and system under multi-dimensional data control provided by the present application with reference to the accompanying drawings.

[0021] Please refer to Figure 1 , which shows a flowchart of the steps of a wildfire evolution risk grading method under multi-dimensional data control provided by an embodiment of the present application. The method includes the following steps: Step 1, divide the area where the transmission line to be analyzed is located into grids, and obtain various fire risk factors of each grid in real time, where there is a wind direction vector.

[0022] Collect the geographical features, vegetation features, and historical fire point data of the area where the transmission line to be analyzed is located from the power department. Among them, the geographical feature factors include 3 factors: altitude, slope, and aspect. The vegetation feature factors include 2 factors: combustible risk level and vegetation index. The combustible risk level includes: non-combustibles, difficult-to-combustibles, ordinary fire-risk materials, and flammable materials. In order to utilize the combustible risk level in subsequent calculations, non-combustibles, difficult-to-combustibles, ordinary fire-risk materials, and flammable materials in the combustible risk level are respectively assigned values of 1, 2, 3, and 4. It should be noted that: 1, 2, 3, and 4 are only an embodiment of this application, and the implementer can set their specific values by themselves.

[0023] To guide the subsequent wildfire prevention work, grid division is carried out on the area where the transmission line to be analyzed is located, and the geographical feature factors and vegetation feature factors are extracted to different grids using geographic information software.

[0024] In this embodiment, the size of the grid is , and the size of the grid can be set by the implementer himself / herself, and this application does not make special restrictions.

[0025] Factors such as temperature, humidity, and light intensity change rapidly over time and have a strong correlation with the wildfire risk. Therefore, a variety of intelligent sensors need to be installed in each grid respectively to achieve real-time monitoring of various micro-environment indicators and improve the timeliness of wildfire risk early warning.

[0026] In this application, a temperature and humidity intelligent sensor, a photosensitive intelligent sensor, a wind speed sensor, a smoke concentration intelligent sensor, a gas concentration intelligent sensor, and a dielectric constant intelligent sensor are installed in each grid respectively. Among them, the temperature and humidity intelligent sensor is used to collect temperature and humidity, the photosensitive intelligent sensor is used to collect light intensity, the wind speed sensor is used to collect wind speed and wind direction, the smoke concentration intelligent sensor is used to collect smoke concentration, the gas concentration intelligent sensor is used to collect gas concentration, and the dielectric constant intelligent sensor is used to collect dielectric constant. Among them, the format of the wind direction data is a wind direction vector with a modulus value of 1. Altitude, slope, aspect, combustible risk level, vegetation index, temperature, humidity, light intensity, wind speed, wind direction, smoke concentration, gas concentration, and dielectric constant are used as each fire risk factor.

[0027] Step 2, for the fire risk factors that have a negative impact on the wildfire risk, perform positive processing to obtain the wildfire spread characteristic value and correction coefficient of each grid, and correct the wildfire spread speed of each grid obtained under the forest fire spread model.

[0028] Since different fire risk factors have different impacts on wildfire risk. For example, the higher the temperature, the higher the wildfire risk; however, the higher the humidity, the lower the wildfire risk. Fire risk factors that have a negative impact on wildfire risk are denoted as negative fire risk factors. For the convenience of subsequent data processing, positive normalization is performed on each negative fire risk factor except the wind direction vector, and the expression is: ; where represents the value of the v-th negative fire risk factor of the j-th grid after positive normalization; represents the maximum value among the original values of the v-th negative fire risk factor of all grids; represents the original value of the v-th negative fire risk factor of the j-th grid.

[0029] Since there are significant differences in the value ranges of different fire risk factors, to improve the usability of the data and the efficiency and quality of monitoring and analysis, after positive normalization of the negative fire risk factors, further normalization is performed on the values of various fire risk factors except the wind direction.

[0030] In this embodiment, the Min - Max normalization method is used to perform normalization on the values of various fire risk factors respectively.

[0031] The wildfire spread speed is one of the important factors reflecting the severity of wildfires and has a great impact on wildfire risk warning. The influence of the superposition of different factors on the wildfire spread speed is not considered in the forest fire spread model, and it only considers the influence of wildfire spread between adjacent grids, lacking an analysis of the wildfire spread process at a larger spatial scale, making it difficult to accurately obtain the wildfire spread risk in a longer time and spatial dimension. Therefore, this application corrects the wildfire spread speed obtained under the forest fire spread model to obtain a more accurate wildfire spread speed.

[0032] Step 2.1, by analyzing the dispersion of various fire risk factors of all grids, obtain the weights of various fire risk factors, and combine the various fire risk factors of each grid to obtain the wildfire spread characteristic values of each grid.

[0033] Taking all fire risk factors except the wind direction of all grids as input, the objective weighting method is used to obtain the weights of each fire risk factor except the wind direction. The greater the weight, the greater the role of each fire risk factor in wildfire spread, and when the values of various fire risk factors of each grid are larger, the wildfire spread risk of the grid is higher.

[0034] In this embodiment, the entropy weight method is used to obtain the weights of each fire risk factor except the wind direction. As other implementation methods, on the basis of being able to obtain the weights of each fire risk factor except the wind direction, implementers can use other existing technologies, such as the standard deviation method, the CRITIC method, etc., and this application does not make special restrictions.

[0035] Taking the j-th grid as an example, with the j-th grid as the center, construct the neighborhood range. Each grid within the neighborhood range except the central grid is used as each neighboring grid of the j-th grid. When the neighborhood range exceeds the area where the transmission line to be analyzed is located, for the exceeded part, a filling method is used for filling.

[0036] It should be noted that: This is only an embodiment of the present application. The implementer can set its specific value by himself / herself, and the present application does not make special restrictions.

[0037] In this embodiment, the mean filling method is used for filling, and the implementer can select other feasible filling methods by himself / herself.

[0038] Through various forest fire risk factors and their weights of each grid, obtain the forest fire spread characteristic value of each grid. The expression is: ; In the formula, represents the forest fire spread characteristic value of the j-th grid; M represents the number of forest fire risk factors except the wind direction; represents the weight of the i-th forest fire risk factor except the wind direction; represents the value of the i-th forest fire risk factor except the wind direction of the j-th grid.

[0039] It should be noted that: The forest fire spread characteristic value is used to reflect the level of the risk of forest fire spread within the grid. The higher the risk of forest fire spread within the grid, the larger the forest fire spread characteristic value.

[0040] Step 2.2, preset each neighboring grid of each grid, obtain the position vector and distance of each neighboring grid relative to each grid. Through the similarity between the wind direction vector of each grid and the position vector, and the forest fire spread characteristic value, and in combination with the distance, obtain the correction coefficient of each grid, and correct the forest fire spread speed of each grid obtained under the forest fire spread model.

[0041] Furthermore, considering the effect of the wind direction on the forest fire spread, still taking the j-th grid as an example, obtain the position vector of each neighboring grid of the j-th grid relative to the j-th grid through the geographic information software, and calculate the similarity between the position vector corresponding to each neighboring grid of the j-th grid and the wind direction vector of the j-th grid, which is used to reflect the influence of the wind direction on the forest fire spread speed between two grids. The greater the similarity, the stronger the effect of the wind direction on enhancing the forest fire spread speed between two grids. The closer the distance between the j-th grid and its neighboring grid, the greater the influence on the forest fire spread. At the same time, the higher the average level of the risk of forest fire spread between the j-th grid and its neighboring grids, the faster the forest fire spreads between them.

[0042] In this embodiment, the similarity between the position vector and the wind direction vector is the cosine similarity. The distance between grids is the Euclidean distance.

[0043] Based on the above analysis, by the similarity between the wind direction vector of each grid and the position vector corresponding to each of its neighboring grids, the average level of the wildfire spread eigenvalue between each grid and each of its neighboring grids, and in combination with the distance between each grid and each of its neighboring grids, the correction coefficient of each grid is obtained, which is used to correct the calculation result of the wildfire spread speed in the wildfire spread model. The specific process is as follows: Still taking the j-th grid as an example, calculate the mean value of the wildfire spread eigenvalues between the j-th grid and each of its neighboring grids; calculate the product of the absolute value of the similarity and the mean value; the correction coefficient of the j-th grid is proportional to the product corresponding to each neighboring grid of the j-th grid and inversely proportional to the distance from the j-th grid to each of its neighboring grids. The expression is: ; in the formula, represents the correction coefficient of the j-th grid; N represents the number of neighboring grids of the j-th grid; represents the product corresponding to the b-th neighboring grid of the j-th grid; ε represents a preset positive number used to avoid the denominator being 0. The value of ε is preset manually and the implementer can set it by himself. In this embodiment, the value of ε is 0.01; norm( ) represents the normalization function; represents the distance between the j-th grid and its b-th neighboring grid. In this embodiment, the Min-Max normalization method is used to normalize the distance.

[0044] Furthermore, through the correction coefficient of each grid, the wildfire spread speed of each grid obtained under the wildfire spread model is corrected. The method is: Take the product of the wildfire spread speed of each grid obtained under the wildfire spread model and the correction coefficient as the corrected wildfire spread speed of each grid.

[0045] Step 3: Use the pre-trained Bayesian classifier to obtain the probability of each grid causing a wildfire and the conditional probability of each value of each fire risk factor in each grid causing a wildfire. Through the correction result, the distance, and the probability of each grid, the characteristic factor of each grid is obtained.

[0046] To accurately grade the wildfire evolution risk of each grid, it is necessary to simultaneously consider the wildfire risk levels of different grids in terms of both wildfire spread and wildfire occurrence.

[0047] Preset a training set and a test set, train a Naive Bayes classifier, input all kinds of fire risk factors of each grid obtained by real-time acquisition into the trained Naive Bayes classifier, and output the probability of wildfire occurrence in each grid and the conditional probability of wildfire occurrence when the values of each fire risk factor in each grid are taken. The training process of the Naive Bayes classifier is a well-known technology and will not be elaborated in this application. It should be noted that: the Naive Bayes classifier is only an embodiment of this application, and the implementer can select other feasible Bayes classifiers by himself.

[0048] In this embodiment, for each grid in the area where the transmission line to be analyzed is located, 100 historical moments are randomly selected, and the values of various fire risk factors of each grid at 100 historical moments and the historical fire point data of each grid at 100 historical moments are respectively obtained. The values of all kinds of fire risk factors and the historical fire point data of each grid at each historical moment are combined into a training vector, and all training vectors are divided into a training set and a test set according to the ratio of 7:3. Among them, both 100 and 7:3 are preset artificially, and the implementer can set them according to the actual situation by himself, and this application does not make special restrictions.

[0049] On the one hand, the higher the probability of wildfire occurrence in the neighboring grids of each grid, the higher the wildfire spread risk of each grid; on the other hand, the closer the distance between each grid and its neighboring grids and the faster the wildfire spread speed of the neighboring grids of each grid, the greater the role of each grid in accelerating the wildfire spread after the wildfire occurs under the current time and space conditions.

[0050] Based on the above analysis, through the corrected wildfire spread speed, the probability of wildfire occurrence of each neighboring grid of each grid, and the distance between each grid and its each neighboring grid, the characteristic factor of each grid is obtained, which is used to characterize the overall level of the wildfire spread risk of each grid after the wildfire occurs in its neighboring grid. The expression is: ; In the formula, represents the characteristic factor of the jth grid; norm( ) represents the normalization function; N represents the number of neighboring grids of the jth grid; represents the corrected wildfire spread speed of the bth neighboring grid of the jth grid; represents the probability of wildfire occurrence of the bth neighboring grid of the jth grid; represents the distance between the jth grid and its bth neighboring grid.

[0051] In this embodiment, the Min-Max normalization method is used to normalize for processing.

[0052] Step 4: Use a comprehensive evaluation method to obtain the stable weights of each fire risk factor for each grid. Based on the characteristic factors, the stable weights, and the conditional probabilities, obtain the risk scores for each grid, and classify the wildfire evolution risks for each grid.

[0053] Furthermore, considering the impact of the interaction between different environmental factors on wildfire ignition, take all fire risk factors of all grids as inputs, and use the Delphi-Analytic Network Process to output the stable weights of various fire risk factors for each grid, reflecting the degree of influence of all fire risk factors in the entire region on the wildfire risks caused by various fire risk factors for each grid. The value range of the stable weights is [0, 1]. Among them, the process of using the Delphi-Analytic Network Process to output the stable weights of various fire risk factors for each grid is a well-known technology, and will not be elaborated in this application. It should be noted that: The Delphi-Analytic Network Process is only an embodiment of this application. Based on the ability to obtain the stable weights of each fire risk factor for each grid, the implementer can select other feasible methods by himself.

[0054] Based on the above analysis, obtain the risk score for each grid through the characteristic factors, the stable weights, and the conditional probabilities. The expression is: ; where represents the risk score of the j-th grid; represents the characteristic factor of the j-th grid; M represents the number of fire risk factors other than wind direction; represents the stable weight of the i-th fire risk factor other than wind direction for the j-th grid; represents the conditional probability of wildfire ignition caused by the value of the i-th fire risk factor other than wind direction for the j-th grid. The schematic diagram of the process for obtaining the risk score is as shown in Figure 2 shown.

[0055] Preset the value ranges for each wildfire evolution risk level. When the risk score of any grid is within the value range of any wildfire evolution risk level, then classify the risk of the any grid as the any wildfire evolution risk level.

[0056] In this embodiment, there are 5 wildfire evolution risk levels, namely no fire risk, low fire risk, medium fire risk, high fire risk, and extremely high fire risk. The value of no fire risk is 0, the value range of low fire risk is (0, 0.25], the value range of medium fire risk is (0.25, 0.5], the value range of high fire risk is (0.5, 0.75], and the value range of extremely high fire risk is (0.75, 1]. Among them, the value ranges of various wildfire evolution risk levels are all preset artificially. The implementer can set the value ranges of various wildfire evolution risk levels according to the actual situation, and this application does not make special restrictions.

[0057] By grading the wildfire evolution risks of different grids in the area where the transmission line to be analyzed is located, it is possible to guide the staff to conduct more frequent inspections on areas with higher risks.

[0058] Based on the same inventive concept as the above method, the embodiment of the present application also provides a wildfire evolution risk grading system under multi-dimensional data control, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above methods for wildfire evolution risk grading under multi-dimensional data control.

[0059] In summary, the present application uses a variety of intelligent sensors to monitor various micro-environment indicators, and positively processes the fire risk factors collected by the intelligent sensors that have a negative impact on the wildfire risk, improving the availability of data and the efficiency and quality of monitoring and analysis; Furthermore, by analyzing the dispersion of each fire risk factor to obtain the weight, and then obtaining the wildfire spread characteristic value, while considering the influence of the wind direction on the wildfire spread, a correction coefficient is obtained to correct the wildfire spread model. Under the superposition effect of a large time and space scale and multiple fire risk factors, the calculation accuracy of the wildfire spread speed is significantly improved, thereby enhancing the reliability of the wildfire evolution risk assessment; Furthermore, while considering the wildfire risks in both aspects of wildfire spread and wildfire occurrence in each grid, a comprehensive evaluation method is adopted to analyze the interaction between different fire risk factors on the wildfire risk, which can more comprehensively reflect the comprehensive wildfire risk of each grid and improve the accuracy of wildfire evolution risk grading.

[0060] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to embodiments of the present disclosure. In this regard, each block in the flowchart or block diagram may represent a module, a segment of code, or a part thereof that contains one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions noted in the blocks may occur in a different order than noted in the accompanying drawings. For example, two consecutive blocks may in fact be executed substantially in parallel, or they may sometimes be executed in the reverse order, depending on the functions involved. In the description corresponding to the flowcharts and block diagrams in the accompanying drawings, the operations or steps corresponding to different blocks may also occur in a different order than disclosed in the description, and sometimes there is no specific order between different operations or steps. For example, two consecutive operations or steps may in fact be executed substantially in parallel, or they may sometimes be executed in the reverse order, depending on the functions involved. Each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented by a dedicated hardware-based system that performs the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0061] For those skilled in the art, it is obvious that the present application is not limited to the details of the above-described exemplary embodiments, and without departing from the basic characteristics of the present application, the present application can be implemented in other specific forms. Therefore, from any point of view, the above embodiments of the present application should be regarded as exemplary and non-limiting.

Claims

1. A method for grading wildfire evolution risk under multi-dimensional data control, characterized in that: The method comprises the following steps: Divide the preset area into grids and obtain various fire risk factors of each grid in real time, including wind direction vectors; For the fire risk factors that have a negative impact on wildfire risk, positive processing is carried out. By analyzing the discreteness of various fire risk factors of all grids, the weights of various fire risk factors are obtained. Combining various fire risk factors of each grid, the wildfire spread characteristic value of each grid is obtained. Preset each neighboring grid of each grid, obtain the position vector and distance of each neighboring grid relative to each grid, obtain the correction coefficient of each grid through the similarity between the wind direction vector of each grid and the position vector, and the wildfire spread characteristic value, and combine the distance, and correct the wildfire spread speed of each grid obtained under the forest fire spread model; A pre-trained Bayesian classifier is used to obtain the probability of each grid causing a wildfire and the conditional probability of each fire risk factor value in each grid causing a wildfire, and the characteristic factor of each grid is obtained through the correction results of each neighboring grid of each grid, the probability, and the distance; A comprehensive evaluation method is used to obtain the stable weights of each fire risk factor of each grid, and the risk score of each grid is obtained through the characteristic factors, the stable weights and the conditional probability, wherein the risk score is used to grade the wildfire evolution risk of each grid.

2. The method for grading wildfire evolution risk under multidimensional data control as claimed in claim 1, characterized in that: The wildfire spread characteristic value is the sum of the products of all fire risk factors of each grid except the wind direction vector and their weights, wherein the weight of each fire risk factor is obtained by an objective weighting method.

3. The method for grading wildfire evolution risk under multidimensional data control as claimed in claim 1, characterized in that: The process of obtaining the correction coefficient is as follows: Calculate the mean of the wildfire spread characteristic values ​​of each grid and its neighboring grids; Calculating the product of the absolute value of the similarity and the mean; The correction coefficient is proportional to the product corresponding to each neighboring grid of each grid, and inversely proportional to the distance.

4. The method for grading wildfire evolution risk under multidimensional data control as claimed in claim 3, characterized in that: The expression of the correction coefficient is: ; In the formula, represents the correction coefficient of the j-th grid; N represents the number of neighboring grids of the j-th grid; represents the product corresponding to the bth neighboring grid of the jth grid; ε represents a preset positive number; norm() represents a normalization function; Represents the distance between the jth grid and its bth neighboring grid.

5. The method for grading wildfire evolution risk under multidimensional data control as claimed in claim 1, characterized in that: The correction of the forest fire spread speed of each grid obtained under the forest fire spread model includes: The product of the wildfire spread speed of each grid obtained under the forest fire spread model and the correction coefficient is used as the corrected wildfire spread speed of each grid.

6. The method for grading wildfire evolution risk under multidimensional data control as claimed in claim 5, characterized in that: The expression of the characteristic factor is: ; In the formula, represents the characteristic factor of the j-th grid; norm() represents the normalization function; N represents the number of neighboring grids of the j-th grid; represents the corrected wildfire spread speed of the bth neighboring grid of the jth grid; represents the probability that the bth neighboring grid of the jth grid will cause a wildfire; Represents the distance between the jth grid and its bth neighboring grid.

7. The method for grading wildfire evolution risk under multidimensional data control as claimed in claim 1, characterized in that: The process of obtaining the risk score is as follows: Calculating a product value of the stability weight and the conditional probability; Calculate the average value of the product values ​​of all fire risk factors except the wind direction vector of each grid; The risk score of each grid is obtained by combining the average value with the characteristic factor.

8. The method for grading wildfire evolution risk under multi-dimensional data control as claimed in claim 7, characterized in that: The risk score is the product of the average value and the characteristic factor.

9. The method for grading wildfire evolution risk under multidimensional data control as claimed in claim 1, characterized in that: The method for grading the wildfire evolution risk of each grid is as follows: A value range of each wildfire evolution risk level is preset, and when the risk score of any grid is within the value range of any wildfire evolution risk level, the risk of any grid is graded as any wildfire evolution risk level.

10. A wildfire evolution risk rating system under multi-dimensional data control, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the method for grading wildfire evolution risk under multidimensional data control as described in any one of claims 1-9 are implemented.

Images