A fire risk assessment method, device, terminal equipment and storage medium

Abstract

The present invention discloses a fire risk assessment method, device, terminal equipment and storage medium, including: dividing the area to be assessed into at least one assessment unit; calculating the fire source distance between the center coordinates of each assessment unit and the fire point coordinates; obtaining the fire characteristic value of each assessment unit according to the fire source distance and the initial characteristic value corresponding to the fire type; respectively using the hierarchical analysis method and the entropy method to obtain the first weight value and the second weight value of the fire index of each assessment unit; based on the random forest model, according to the fire index and the fire characteristic value, obtaining the contribution value of the fire index; according to the first weight value, the second weight value and the contribution value, obtaining the third weight value of the fire index; according to the fire index and the third weight value, obtaining the fire risk assessment value of each assessment unit. The embodiment of the present invention can greatly improve the accuracy of the fire risk assessment result.

Description

Technical Field

[0001] The present invention relates to the field of risk analysis and assessment, and in particular to a fire risk assessment method, device, terminal equipment and computer-readable storage medium. Background Art

[0002] With the rapid development of economy, the scale of cities is expanding, various production and living factors and population are concentrating, and various risks are highly concentrated, which has greatly increased the total amount and difficulty of urban firefighting work, resulting in essential changes. Under the original management mode and management mechanism, even if the management intensity is increased, its management efficiency is difficult to be significantly improved. Therefore, it is necessary to propose a fire risk assessment method to guide the optimal deployment of the combination of urban regional fire risk level and fire emergency rescue force, and provide a scientific basis for multi-dimensional judgment and command and disposal in fire prevention and control, fire fighting and rescue.

[0003] However, the existing urban area fire risk assessment mainly considers factors such as fire hazard sources, population density, building characteristics, fire protection facilities and disaster resistance, and uses a weighted average method to evaluate each unit to obtain a fire risk assessment result. This method does not take into account the statistical results of historical fire data, and it is difficult to accurately judge the fire risk based on the fire characteristics of each area in the city. In addition, since the determination of indicator weights in the traditional fire risk assessment system is greatly affected by the subjective experience of decision makers, it is difficult to eliminate the impact of subjective factors on the fire assessment results, thereby reducing the accuracy of the fire assessment results. Summary of the invention

[0004] In view of the above problems, the purpose of the embodiments of the present invention is to provide a fire risk assessment method, apparatus, terminal device and computer-readable storage medium, which can perform feature analysis on regional fire history data and evaluate regional fire risks in combination with hierarchical analysis method, entropy method and random forest algorithm, thereby greatly improving the accuracy of fire risk assessment results.

[0005] In order to achieve the above-mentioned object, a first aspect of an embodiment of the present invention provides a fire risk assessment method, comprising: dividing an area to be assessed into at least one assessment unit; obtaining a fire point coordinate M of a historical fire area of ​​each assessment unit; jk , and calculate the center coordinates N of each evaluation unit j Coordinate with the fire point M jk The distance between fire sources, M jk represents the coordinates of the kth fire point in the jth evaluation unit, N jRepresents the center coordinates of the jth evaluation unit; obtains the fire characteristic value of each evaluation unit according to the initial characteristic values ​​corresponding to the fire source distance and the fire type; adopts the hierarchical analysis method to obtain the first weight value of the fire index of each evaluation unit; adopts the entropy method to obtain the second weight value of the fire index; based on the random forest model, obtains the contribution value of the fire index according to the fire index and the fire characteristic value; obtains the third weight value of the fire index according to the first weight value, the second weight value and the contribution value; obtains the fire risk assessment value of each evaluation unit according to the fire index and the third weight value.

[0006] Furthermore, the fire risk assessment method obtains the random forest model through the following steps: quantifying the fire index to obtain the quantized value of the fire index; using the quantized value of the fire index as a feature value and the fire feature value as a label value, and using a random forest algorithm to generate a decision tree to obtain the random forest model.

[0007] Furthermore, based on the random forest model, the contribution value of the fire index is obtained according to the fire index and the fire characteristic value, specifically: using the random forest algorithm, obtaining the first out-of-bag data error of the decision tree according to the first out-of-bag data corresponding to the characteristic value; adding noise interference to the first out-of-bag data to obtain the second out-of-bag data; using the random forest algorithm, obtaining the second out-of-bag data error of the decision tree according to the second out-of-bag data; obtaining the contribution value of the fire index according to the first out-of-bag data error, the second out-of-bag data and the number of decision trees.

[0008] Furthermore, the acquisition of the ignition point coordinates M of the historical fire area of ​​each evaluation unit jk , and calculate the center coordinates N of each evaluation unit j Coordinate with the fire point M jk Specifically, for the kth fire point coordinates in the jth evaluation unit, the map Web service API is called to obtain the geographic coordinates of the historical fire area and the geographic coordinates of the center of the evaluation unit; based on ARCMap, the projection method is used to obtain the fire point coordinates M of the historical fire area. jk and the center coordinate N of the evaluation unit j ;according to Calculate the fire source distance; where r represents the fire source distance, x represents the horizontal coordinate of the fire point coordinate, y represents the vertical coordinate of the fire point coordinate, and x i The horizontal coordinate representing the center coordinate, y i The vertical coordinate representing the center coordinate.

[0009] Furthermore, the fire characteristic value of each evaluation unit is obtained according to the initial characteristic value corresponding to the fire source distance and the fire type, specifically: the fire characteristic value of each fire is obtained according to the following formula:

[0010]

[0011] Among them, F represents the fire characteristic value of the fire, F0 represents the initial characteristic value corresponding to the fire type, r represents the distance from the fire source, R1 represents the inner radius of the fire impact range, and R2 represents the outer radius of the fire impact range; the fire characteristic values ​​of each fire are superimposed to obtain the fire characteristic value of each evaluation unit.

[0012] Further, the third weight value of the fire index is obtained according to the first weight value, the second weight value and the contribution value, specifically: according to Obtain correction coefficient; according to W i =(ε×α i +(1-ε)×β i ) obtains the third weight value of the fire index; wherein n represents the number of indicators, Y i Represents the deviation value of the random forest algorithm, α i Indicates the first weight value, β i represents the second weight value, θ i Indicates the contribution value, ε means the correction coefficient, W i Indicates the third weight value.

[0013] Furthermore, the fire risk assessment value of each of the assessment units is obtained based on the fire index and the third weight value, specifically: the fire index is quantized to obtain the quantized value of the fire index; the quantized value of the fire index is multiplied by the third weight value to obtain a child index assessment value; the assessment values ​​of the child indexes belonging to the same parent index are added up to obtain a parent index assessment value; the parent index assessment values ​​are added up to obtain a fire risk assessment value of each of the assessment units.

[0014] A second aspect of an embodiment of the present invention provides a fire risk assessment device, comprising an assessment unit division module, for dividing a to-be-assessed area into at least one assessment unit; a fire source distance acquisition module, for acquiring a fire point coordinate M of a historical fire area of ​​each of the assessment units; jk , and calculate the center coordinates N of each evaluation unit j Coordinate with the fire point M jk The distance between fire sources, M jk represents the coordinates of the kth fire point in the jth evaluation unit, N jRepresents the center coordinates of the jth evaluation unit; a fire characteristic value acquisition module, used to obtain the fire characteristic value of each of the evaluation units according to the initial characteristic value corresponding to the fire source distance and the fire type; a first weight value acquisition module, used to obtain the first weight value of the fire index of each of the evaluation units by adopting the hierarchical analysis method; a second weight value acquisition module, used to obtain the second weight value of the fire index by adopting the entropy method; a contribution value acquisition module, used to obtain the contribution value of the fire index according to the fire index and the fire characteristic value based on a random forest model; a third weight value acquisition module, used to obtain the third weight value of the fire index according to the first weight value, the second weight value and the contribution value; a fire risk assessment module, used to obtain the fire risk assessment value of each of the evaluation units according to the fire index and the third weight value.

[0015] A third aspect of an embodiment of the present invention provides a terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the fire risk assessment method as described in any one of the first aspects when executing the computer program.

[0016] A fourth aspect of an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by one or more processors, the fire risk assessment method as described in any one of the first aspects is implemented.

[0017] Compared with the prior art, the beneficial effects of the embodiments of the present invention are as follows: a fire risk assessment method, apparatus, terminal device and computer-readable storage medium provided by the embodiments of the present invention divide the area into units, perform feature analysis on the fire history data of each unit, output the indicator weights of the comprehensive analytic hierarchy process and the entropy method, and use the random forest algorithm to correct the indicator weight values ​​to form objective and accurate indicator weight values, and assess the regional fire risk based on the indicator weight values, thereby greatly improving the accuracy of the fire risk assessment results and accurately reflecting the spatial distribution characteristics of the regional fire risk. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a flow chart of a preferred embodiment of a fire risk assessment method provided by the present invention;

[0019] Figure 2 It is a schematic diagram of the index hierarchy structure of a preferred embodiment of the hierarchical analysis method provided by the present invention;

[0020] Figure 3 It is a structural schematic diagram of a preferred embodiment of a fire risk assessment device provided by the present invention. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] See also Figure 1 , is a flow chart of a preferred embodiment of a fire risk assessment method provided by the present invention.

[0023] A first aspect of an embodiment of the present invention provides a fire risk assessment method, including steps S1 to S9, which are specifically as follows:

[0024] Step S1: Divide the area to be evaluated into at least one evaluation unit.

[0025] Preferably, in the GIS system, a fishnet creation tool is used to divide the assessment area into several independent assessment units with a grid size of 500m×500m.

[0026] Step S2: Obtain the ignition point coordinates M of the historical fire area of ​​each evaluation unit jk , and calculate the center coordinates N of each evaluation unit j Coordinate with the fire point M jk The distance between fire sources, M jk represents the coordinates of the kth fire point in the jth evaluation unit, N j Represents the center coordinate of the jth evaluation unit.

[0027] Wherein, the ignition point coordinates M of the historical fire area of ​​each evaluation unit are obtained. jk , and calculate the center coordinates N of each evaluation unit j Coordinate with the fire point M jk The distance between fire sources, M jk represents the coordinates of the kth fire point in the jth evaluation unit, N j represents the center coordinates of the jth evaluation unit, specifically:

[0028] For the kth fire point coordinates in the jth evaluation unit, call the map Web service API to obtain the geographic coordinates of the historical fire area and the geographic coordinates of the center of the evaluation unit;

[0029] Based on ARCMap, the projection method is used to obtain the fire point coordinates M in the historical fire area. jk and the center coordinate N of the evaluation unitj ;

[0030] according to The fire source distance can be obtained;

[0031] Among them, r represents the distance from the fire source, x represents the horizontal coordinate of the fire point coordinate, y represents the vertical coordinate of the fire point coordinate, and x i The horizontal coordinate representing the center coordinate, y i The vertical coordinate representing the center coordinate.

[0032] It should be noted that the map Web service API may use the Web service API of map software such as Baidu Maps, Amap, etc., and no specific limitation is made here.

[0033] Step S3: Obtaining the fire characteristic value of each evaluation unit according to the initial characteristic value corresponding to the fire source distance and the fire type.

[0034] The fire characteristic value of each evaluation unit is obtained according to the initial characteristic value corresponding to the fire source distance and the fire type, specifically:

[0035] The fire characteristic value of each fire is obtained according to the following formula:

[0036]

[0037] Among them, F represents the fire characteristic value of the fire, F0 represents the initial characteristic value corresponding to the fire type, r represents the distance from the fire source, R1 represents the inner radius of the fire impact range, and R2 represents the outer radius of the fire impact range;

[0038] The fire characteristic values ​​of each fire are superimposed to obtain the fire characteristic value of each evaluation unit.

[0039] It can be seen that when r≤R1, the characteristic value of the fire caused by the fire source remains unchanged, which is called the characteristic invariant zone; when R1≤r≤R2, the characteristic value of the fire caused by the fire source decreases with the increase of distance, and the rate of decrease becomes faster and faster until it approaches 0, which is called the characteristic gradient zone.

[0040] It should be noted that, according to the different causes of fire, the initial characteristic values ​​and radiation radius of the fire are shown in Table 1:

[0041] Table 1 Fire initial characteristic values ​​and radiation radius of different fire types

[0042]

[0043]

[0044] Step S4: using the analytic hierarchy process to obtain the first weight value of the fire index of each evaluation unit.

[0045] Specifically, the AHP method first constructs a hierarchical indicator system based on the factors and factor sub-levels that affect regional fire risks, such as Figure 2 After the hierarchical indicator system is established, the subordinate relationship between the indicators of the upper and lower levels is determined. Assuming that the target of an indicator at a certain level is A k , the next level of criteria it governs are B1, B2, ... B n , then assign B1, B2, ...B according to their relative importance to the target. n The corresponding weight value is B1, B2,...B n For Target A k When the relative importance of can be directly expressed quantitatively, their corresponding weights can be determined. The method used by the hierarchical analysis method to determine the weights is the pairwise comparison method. The comparison results are written into the judgment matrix in the form of numbers to obtain the judgment matrix of the criterion layer B for the target layer A, as shown in Table 2:

[0046] Table 2 Judgment matrix

[0047] <![CDATA[A k ]]> <![CDATA[B1]]> <![CDATA[B2]]> ... <![CDATA[B j ]]> <![CDATA[B1]]> <![CDATA[b 11 ]]> <![CDATA[b 12 ]]> ... <![CDATA[b 1j ]]> <![CDATA[B2]]> <![CDATA[b 21 ]]> <![CDATA[b 22 ]]> ... <![CDATA[b 2j ]]> ... ... ... ... ... <![CDATA[B i ]]> <![CDATA[b i1 ]]> <![CDATA[b i2 ]]> ... <![CDATA[b ij ]]>

[0048] In Table 2, b ij =B i / B j , means for A k This evaluation indicator, factor B i Factor B j Numerical expression of relative importance judgment. b ij The value of can be expressed as 1-9 and its reciprocal as a relative importance scale, as shown in Table 3:

[0049] Table 3 Relative importance scale and its meaning

[0050]

[0051]

[0052] Using the scale of the above judgment matrix, factor B i and factor B j Comparison of two to find b ij , and then form a quantitative judgment matrix B. Find the maximum eigenvalue root of the judgment matrix B and its corresponding eigenvector, normalize the maximum eigenvector corresponding to the maximum eigenvalue root, and get the criteria B1, B2, ... B nThe weight value of target A. The calculation of the composite weight should be from bottom to top, multiplying the weight value of each indicator in the last layer with the relative weight value of the previous layer in turn, so as to form the composite weight of each indicator for the total target. Define the weight of each indicator in the hierarchical analysis method as α i .

[0053] Step S5: using an entropy method to obtain a second weight value of the fire index.

[0054] Specifically, the first step of the entropy method is to standardize the data matrix. Assuming there are m indicators and n evaluation units, the original matrix is:

[0055]

[0056] Perform standardization:

[0057]

[0058] Among them, r ab is the standard value of the b-th indicator on the a-th evaluation unit.

[0059] Then, the standardized data matrix is:

[0060] R=(r ab ) m×n

[0061] The second step of the entropy method is to calculate the entropy value of each indicator and calculate the weight P of the bth indicator to the ath evaluation unit. ab for:

[0062]

[0063] Then we obtain the entropy value e of the b-th index b for:

[0064]

[0065] in,

[0066] Then we obtain the difference coefficient g of the b-th index b for:

[0067] g b =1-e b

[0068] For b-th indicator, X ab The greater the difference, the greater the impact on the evaluation results and the smaller the entropy value.

[0069] The third step of the entropy method is to calculate the weight of each indicator. The weight of the bth indicator is:

[0070]

[0071] Step S6: Based on the random forest model, the contribution value of the fire index is obtained according to the fire index and the fire characteristic value.

[0072] Among them, the fire risk assessment method obtains the random forest model through the following steps: quantifying the fire index to obtain the quantized value of the fire index; using the quantized value of the fire index as the feature value and the fire feature value as the label value, and using the random forest algorithm to generate a decision tree to obtain the random forest model.

[0073] Preferably, the fire index is quantified by the following method:

[0074] For dimensional indicators, the percentage of the service area within each assessment unit can be used as the quantitative indicator value; the quantile method can also be used to divide the fire data within the assessment unit into equal parts for quantification.

[0075] For dimensionless indicators, they are qualitatively and quantitatively processed according to their risk levels and corresponding quantitative ranges.

[0076] Specifically, random forest is an integrated algorithm. The main idea is to vote on the classification results of several weak classifiers to form a strong classifier. "Random" is mainly reflected in random sampling and random feature selection, which effectively prevents overfitting; "Forest" is reflected in the generation of multiple decision trees, which prevents the occurrence of low model generalization ability and has excellent big data processing capabilities.

[0077] A decision tree is a tree structure that classifies and predicts new data by measuring historical data. A decision tree consists of three main parts: decision nodes, branches, and leaf nodes. The top decision node of the decision tree is the root decision node, and each branch has a new decision node. Below the decision node are leaf nodes, each of which represents an indicator value to be classified, and each leaf node represents a result. The entire decision process starts from the root decision node, and from top to bottom, different results are given at each decision node according to the classification of the data.

[0078] The decision tree is constructed using the greedy idea, that is, the attribute that can get the best split result is selected for splitting. In order to get the best split result, the data of the child node should be as "pure" as possible after each split, and the basis for selection is the information gain rate:

[0079]

[0080] Among them, Gain represents the complexity of the node. The higher the Gain, the higher the complexity. Information gain is the sum of the data complexity before splitting minus the data complexity of the child nodes. The greater the information gain, the more the complexity is reduced after splitting, and the more obvious the classification effect.

[0081] The complexity of a node can be calculated in two different ways: entropy and Gini coefficient. The embodiment of the present invention determines the optimal attribute by using the Gini coefficient:

[0082]

[0083] Among them, p i is the proportion of the number of class i. After selecting the decision tree information splitting attribute, because it cannot be allowed to grow indefinitely, a minimum number of nodes is proposed, that is, when the amount of data of the node is less than a specified number, the split will not continue. There are two reasons why the decision tree cannot be allowed to grow indefinitely: first, when the amount of data is small, further splitting is likely to strengthen the effect of noise data; second, to reduce the complexity of tree growth, ending the split in advance is helpful to reduce the impact of overfitting to a certain extent.

[0084] Furthermore, based on the random forest model, the contribution value of the fire index is obtained according to the fire index and the fire characteristic value, specifically: using the random forest algorithm, obtaining the first out-of-bag data error of the decision tree according to the first out-of-bag data corresponding to the characteristic value; adding noise interference to the first out-of-bag data to obtain the second out-of-bag data; using the random forest algorithm, obtaining the second out-of-bag data error of the decision tree according to the second out-of-bag data; obtaining the contribution value of the fire index according to the first out-of-bag data error, the second out-of-bag data and the number of decision trees.

[0085] Specifically, for each decision tree, the corresponding out-of-bag data (OOB) is selected to calculate the out-of-bag data error, which is recorded as errOOB1. The out-of-bag data refers to a set of data obtained by repeatedly sampling samples and feature values ​​each time a decision tree is established in step S6 for training the decision tree. At this time, about 1 / 3 of the data is not used, that is, it does not participate in the establishment of the decision tree. This part of the unused data can be used to evaluate the performance of the decision tree and calculate the prediction error rate of the model, which is the out-of-bag data error.

[0086] Noise interference is randomly added to the fire features corresponding to all samples of the out-of-bag data. The embodiment of the present invention adopts a method of randomly changing the value of the sample at its corresponding fire feature to introduce noise interference, and the out-of-bag data error is calculated again, which is recorded as errOOB2.

[0087] Assuming there are N decision trees in the random forest model, the contribution value of each fire feature θi for: The reason why this value can illustrate the importance of fire characteristics is that if random noise interference is added, the accuracy of the out-of-bag data will drop significantly (that is, errOOB2 will increase), which means that this fire feature has a great impact on the prediction results of the sample, and thus shows that its importance is relatively high.

[0088] It should be noted that the random forest algorithm used in the embodiment of the present invention is implemented based on the scikit-learn package in Python. In addition, other platforms can also be used to implement the random forest algorithm, which is not specifically limited here.

[0089] Step S7: Obtain a third weight value of the fire index according to the first weight value, the second weight value and the contribution value.

[0090] The method of obtaining the third weight value of the fire index according to the first weight value, the second weight value and the contribution value is specifically as follows:

[0091] according to Obtain correction coefficients;

[0092] According to W i =(ε×α i +(1-ε)×β i ) Obtain the third weight value of the fire indicator;

[0093] where n represents the number of indicators, Y i Represents the deviation value of the random forest algorithm, α i Indicates the first weight value, β i represents the second weight value, θ i Indicates the contribution value, ε means the correction coefficient, W i Indicates the third weight value.

[0094] It should be noted that the deviation value Y of each indicator for the random forest algorithm i =|W i -θ i |, when When it is minimum, the obtained ε is the final correction coefficient, thereby obtaining the final weight value of each indicator.

[0095] Step S8: Obtaining a fire risk assessment value of each of the assessment units according to the fire index and the third weight value.

[0096] The fire risk assessment value of each assessment unit is obtained according to the fire index and the third weight value, specifically:

[0097] quantifying the fire index to obtain a quantified value of the fire index;

[0098] Multiplying the quantized value of the fire index by the third weight value to obtain a sub-level index evaluation value;

[0099] Accumulate the evaluation values ​​of the child indicators belonging to the same parent indicator to obtain the parent indicator evaluation value;

[0100] The parent index evaluation values ​​are accumulated to obtain the fire risk evaluation value of each evaluation unit.

[0101] It should be noted that, based on the set quantitative range and combined with the fire accident grade classification standard, the fire risk is divided into five levels, as shown in Table 4:

[0102] Table 4 Fire risk level classification

[0103]

[0104]

[0105] During the fire risk assessment process, each fire indicator can be classified according to this level standard to achieve uniformity of assessment results. The corresponding scores for each level are as follows: Level I, 80 to 100 points; Level II, 60 to 79 points; Level III, 40 to 59 points; Level IV, 20 to 39 points; Level V, 0 to 19 points.

[0106] A fire risk assessment method provided by an embodiment of the present invention divides a region into units, performs feature analysis on the fire history data of each unit, outputs the indicator weights of a comprehensive hierarchical analysis method and an entropy method, and uses a random forest algorithm to correct the indicator weight values ​​to form objective and accurate indicator weight values. The regional fire risk is assessed based on the indicator weight values, thereby greatly improving the accuracy of the fire risk assessment results and accurately reflecting the spatial distribution characteristics of the regional fire risk.

[0107] See also Figure 3 A second aspect of an embodiment of the present invention provides a fire risk assessment device, comprising:

[0108] An evaluation unit division module 301 is used to divide the area to be evaluated into at least one evaluation unit;

[0109] The fire source distance acquisition module 302 is used to obtain the fire point coordinates M of the historical fire area of ​​each evaluation unit. jk , and calculate the center coordinates N of each evaluation unit j Coordinate with the fire point M jk The distance between fire sources, M jkrepresents the coordinates of the kth fire point in the jth evaluation unit, N j Represents the center coordinates of the jth evaluation unit;

[0110] A fire characteristic value acquisition module 303, configured to obtain a fire characteristic value of each of the evaluation units according to the initial characteristic values ​​corresponding to the fire source distance and the fire type;

[0111] A first weight value acquisition module 304, used to obtain a first weight value of the fire index of each evaluation unit by using a hierarchical analysis method;

[0112] A second weight value acquisition module 305, used to obtain a second weight value of the fire indicator by using an entropy method;

[0113] A contribution value acquisition module 306, configured to obtain a contribution value of the fire index according to the fire index and the fire characteristic value based on a random forest model;

[0114] A third weight value acquisition module 307, configured to obtain a third weight value of the fire index according to the first weight value, the second weight value and the contribution value;

[0115] The fire risk assessment module 308 is used to obtain a fire risk assessment value of each assessment unit according to the fire index and the third weight value.

[0116] Furthermore, the fire source distance acquisition module 302 is also used to: for the kth fire point coordinates in the jth evaluation unit, call the map Web service API to obtain the geographical coordinates of the historical fire area and the geographical coordinates of the center of the evaluation unit; based on ARCMap, use the projection method to obtain the fire point coordinates M of the historical fire area jk and the center coordinate N of the evaluation unit j ;according to Calculate the fire source distance; where r represents the fire source distance, x represents the horizontal coordinate of the fire point coordinate, y represents the vertical coordinate of the fire point coordinate, and x i The horizontal coordinate representing the center coordinate, y i The vertical coordinate representing the center coordinate.

[0117] Furthermore, the fire characteristic value acquisition module 303 is further used to obtain the fire characteristic value of each fire according to the following formula:

[0118]

[0119] Among them, F represents the fire characteristic value of the fire, F0 represents the initial characteristic value corresponding to the fire type, r represents the distance from the fire source, R1 represents the inner radius of the fire impact range, and R2 represents the outer radius of the fire impact range; the fire characteristic values ​​of each fire are superimposed to obtain the fire characteristic value of each evaluation unit.

[0120] Furthermore, the fire risk assessment device also includes a random forest acquisition module 309, which is used to quantify the fire index to obtain the quantized value of the fire index; use the quantized value of the fire index as a feature value, and the fire feature value as a label value, and use a random forest algorithm to generate a decision tree to obtain the random forest model.

[0121] Furthermore, the contribution value acquisition module 306 is also used to: adopt the random forest algorithm to obtain a first out-of-bag data error of the decision tree according to the first out-of-bag data corresponding to the eigenvalue; add noise interference to the first out-of-bag data to obtain second out-of-bag data; adopt the random forest algorithm to obtain a second out-of-bag data error of the decision tree according to the second out-of-bag data; and obtain the contribution value of the fire index according to the first out-of-bag data error, the second out-of-bag data and the number of decision trees.

[0122] Furthermore, the third weight value acquisition module 307 is also used to: Obtain correction coefficient; according to W i =(ε×α i +(1-ε)×β i ) obtains the third weight value of the fire index; wherein n represents the number of indicators, Y i Represents the deviation value of the random forest algorithm, α i Indicates the first weight value, β i represents the second weight value, θ i represents the contribution value, ε represents the correction coefficient, W i Indicates the third weight value.

[0123] Furthermore, the fire risk assessment module 308 is also used to: quantify the fire indicator to obtain a quantized value of the fire indicator; multiply the quantized value of the fire indicator by the third weight value to obtain a child indicator evaluation value; accumulate the child indicator evaluation values ​​belonging to the same parent indicator to obtain a parent indicator evaluation value; accumulate the parent indicator evaluation values ​​to obtain a fire risk assessment value of each of the assessment units.

[0124] It should be noted that a fire risk assessment device provided in an embodiment of the present invention can implement all processes of the fire risk assessment method described in any of the above embodiments, and the functions of each module in the device and the technical effects achieved are respectively the same as the functions and technical effects achieved by the fire risk assessment method described in the above embodiments, and will not be repeated here.

[0125] A third aspect of an embodiment of the present invention provides a terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the fire risk assessment method as described in any embodiment of the first aspect is implemented.

[0126] The terminal device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory. The terminal device may also include an input / output device, a network access device, a bus, etc.

[0127] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the terminal device, and uses various interfaces and lines to connect various parts of the entire terminal device.

[0128] The memory can be used to store the computer program and / or module, and the processor realizes various functions of the terminal device by running or executing the computer program and / or module stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0129] A fourth aspect of an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by one or more processors, the fire risk assessment method as described in any embodiment of the first aspect is implemented.

[0130] Through the description of the above implementation methods, those skilled in the art can clearly understand that the present invention can be implemented by means of software plus a necessary hardware platform, and of course can also be implemented entirely by hardware. Based on such an understanding, all or part of the contribution of the technical solution of the present invention to the background technology can be embodied in the form of a software product, and the computer software product can be stored in a storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., including a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments of the present invention or certain parts of the embodiments.

[0131] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A fire risk assessment method, characterized in that: include: Dividing the area to be evaluated into at least one evaluation unit; Obtain the ignition point coordinates M of each historical fire area of ​​the evaluation unit jk , and calculate the center coordinates N of each evaluation unit j With the fire point coordinates M jk Distance between fire sources, M jk represents the coordinates of the kth fire point in the jth evaluation unit, N j represents the center coordinates of the jth evaluation unit; Obtaining a fire characteristic value of each of the evaluation units according to the initial characteristic values ​​corresponding to the fire source distance and the fire type; Using the hierarchical analysis method to obtain a first weight value of the fire index of each of the evaluation units; Obtaining a second weight value of the fire index by adopting an entropy method; Based on the random forest model, obtaining the contribution value of the fire index according to the fire index and the fire characteristic value; Obtaining a third weight value of the fire index according to the first weight value, the second weight value and the contribution value; Obtaining a fire risk assessment value of each of the assessment units according to the fire index and the third weight value; The fire characteristic value of each evaluation unit is obtained according to the initial characteristic value corresponding to the fire source distance and the fire type, specifically: The fire characteristic value of each fire is obtained according to the following formula: Among them, F represents the fire characteristic value of the fire, F0 represents the initial characteristic value corresponding to the fire type, r represents the distance from the fire source, R1 represents the inner radius of the fire impact range, and R2 represents the outer radius of the fire impact range; The fire characteristic values ​​of each fire are superimposed to obtain the fire characteristic value of each evaluation unit.

2. The fire risk assessment method according to claim 1, characterized in that: The method obtains the random forest model by the following steps: quantifying the fire index to obtain a quantified value of the fire index; The quantified value of the fire index is used as a characteristic value, the fire characteristic value is used as a label value, and a random forest algorithm is used to generate a decision tree to obtain the random forest model.

3. The fire risk assessment method according to claim 2, characterized in that: The random forest model is based on the fire index and the fire characteristic value to obtain the contribution value of the fire index, specifically: Using the random forest algorithm, obtaining a first out-of-bag data error of the decision tree according to the first out-of-bag data corresponding to the eigenvalue; Adding noise interference to the first out-of-bag data to obtain second out-of-bag data; Using the random forest algorithm, obtaining a second out-of-bag data error of the decision tree according to the second out-of-bag data; A contribution value of the fire index is obtained according to the first out-of-bag data error, the second out-of-bag data and the number of decision trees.

4. The fire risk assessment method according to claim 1, characterized in that: The acquisition of the ignition point coordinates M of the historical fire area of ​​each evaluation unit jk , and calculate the center coordinates N of each evaluation unit j With the fire point coordinates M jk The fire source distance between them is: For the kth fire point coordinates in the jth evaluation unit, call the map Web service API to obtain the geographic coordinates of the historical fire area and the geographic coordinates of the center of the evaluation unit; Based on ARCMap, the projection method is used to obtain the fire point coordinates M in the historical fire area. jk and the center coordinates N of the evaluation unit j ; according to Calculate and obtain the fire source distance; Among them, r represents the distance from the fire source, x represents the horizontal coordinate of the fire point coordinate, y represents the vertical coordinate of the fire point coordinate, and x i The horizontal coordinate of the center, y i The vertical coordinate representing the center coordinate.

5. The fire risk assessment method according to claim 1, characterized in that: The step of obtaining the third weight value of the fire index according to the first weight value, the second weight value and the contribution value is specifically: according to Obtain correction factor; According to W i =(ε×α i +(1-ε)×β i ) obtaining a third weight value of the fire index; Among them, n represents the number of indicators, Y i Represents the bias value of the random forest algorithm, α i represents the first weight value, β i represents the second weight value, θ i represents the contribution value, ε represents the correction coefficient, W i Represents the third weight value.

6. The fire risk assessment method according to claim 1, characterized in that: The fire risk assessment value of each assessment unit is obtained according to the fire index and the third weight value, specifically: quantifying the fire index to obtain a quantified value of the fire index; Multiplying the quantized value of the fire index by the third weight value to obtain a sub-level index evaluation value; Accumulate the evaluation values ​​of the child indicators belonging to the same parent indicator to obtain the parent indicator evaluation value; The parent index evaluation values ​​are accumulated to obtain the fire risk evaluation value of each evaluation unit.

7. A fire risk assessment device, characterized in that: include: An evaluation unit division module, used to divide the area to be evaluated into at least one evaluation unit; The fire source distance acquisition module is used to obtain the fire point coordinates M of the historical fire area of ​​each evaluation unit. jk , and calculate the center coordinates N of each evaluation unit j With the fire point coordinates M jk Distance between fire sources, M jk represents the coordinates of the kth fire point in the jth evaluation unit, N j represents the center coordinates of the jth evaluation unit; A fire characteristic value acquisition module, used to obtain the fire characteristic value of each evaluation unit according to the initial characteristic value corresponding to the fire source distance and the fire type; A first weight value acquisition module, used for obtaining a first weight value of the fire index of each evaluation unit by using a hierarchical analysis method; A second weight value acquisition module, used for obtaining a second weight value of the fire indicator by adopting an entropy method; A contribution value acquisition module, used to obtain the contribution value of the fire index according to the fire index and the fire characteristic value based on a random forest model; A third weight value acquisition module, used to obtain a third weight value of the fire index according to the first weight value, the second weight value and the contribution value; A fire risk assessment module, used to obtain a fire risk assessment value of each assessment unit according to the fire index and the third weight value; The fire characteristic value acquisition module is further used to obtain the fire characteristic value of each fire according to the following formula: Among them, F represents the fire characteristic value of the fire, F0 represents the initial characteristic value corresponding to the fire type, r represents the distance from the fire source, R1 represents the inner radius of the fire impact range, and R2 represents the outer radius of the fire impact range; the fire characteristic values ​​of each fire are superimposed to obtain the fire characteristic value of each evaluation unit.

8. A terminal device, characterized in that: The method comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the fire risk assessment method according to any one of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by one or more processors, the fire risk assessment method according to any one of claims 1 to 6 is implemented.

Images