Forest fire barrier knowledge graph construction method and system

Abstract

The invention relates to the technical field of forest fire prevention, in particular to a forest fire barrier knowledge graph construction method and system, and the method comprises the following steps: S1, collecting multi-source data related to forest fire prevention; s2, preprocessing the multi-source data to generate a forest fire barrier data set; s3, extracting knowledge elements about the forest fire prevention isolation belt from the forest fire prevention isolation belt data set, and calibrating association relationships among the elements; s4, on the basis of the knowledge elements extracted in the S3 and the association relationship thereof, constructing a forest fire barrier knowledge graph; and S5, optimizing the knowledge graph according to historical fire data and actual fire prevention experience. According to the invention, through the multi-source data fusion and knowledge graph optimization technology, forest fire prevention related information can be accurately integrated and analyzed, and the scientificity and adaptability of fire barrier planning are improved, so that accurate fire prevention decision and management are effectively supported.

Description

Technical Field

[0001] The present invention relates to the field of forest fire prevention technology, and in particular to a method and system for constructing a knowledge graph of a forest fire isolation zone. Background Art

[0002] With global climate change and frequent forest fires, forest fire prevention has become increasingly complex. Traditional forest fire prevention relies on manual monitoring and ground patrols, and is unable to acquire and process large amounts of multidimensional data related to forest fire prevention in real time. Especially in complex forest environments, factors such as the geographical distribution of forests, meteorological conditions, historical fire data, and vegetation types have a highly nonlinear and complex impact on the spread of fire. Existing fire isolation zone planning is often based on a single factor or limited data sources, which makes it difficult to evaluate and optimize its fire prevention effects, and it is impossible to make accurate fire prevention decisions for different fire environments.

[0003] In the existing technology, although there are some methods that attempt to assist decision-making and planning by constructing knowledge graphs, these methods have significant limitations in practical applications due to the lack of effective integration of multi-source data and insufficient mining of complex relationships in the field of forest fire prevention. For example, traditional knowledge graph construction methods often ignore the inherent connections between multi-dimensional data, resulting in the difficulty of the graph construction process to cover comprehensive information and accurately reflect the complex relationship between fire spread and factors such as geography, meteorology, and vegetation. In addition, most of the existing knowledge graph optimization methods lack a dynamic adjustment mechanism based on historical fire data and actual fire prevention experience, and cannot update and eliminate redundant information in real time. Therefore, how to integrate multi-source data, mine complex associations, and optimize the knowledge graph in real time to construct a knowledge graph of forest fire isolation zones with higher accuracy and practicality is a key technical problem that needs to be solved. Summary of the invention

[0004] Based on the above objectives, the present invention provides a method and system for constructing a knowledge graph of forest fire isolation zones.

[0005] A method for constructing a knowledge graph of a forest fire isolation zone comprises the following steps:

[0006] S1: Collect multi-source data related to forest fire prevention, including forest geographical distribution data, meteorological conditions data, historical fire data and vegetation type distribution data;

[0007] S2: De-noise, standardize, fill in missing values ​​and align time for the multi-source data collected in S1 to generate a forest fire isolation zone dataset;

[0008] S3: Extract knowledge elements about forest fire isolation belts from the forest fire isolation belt dataset, including geographical features, meteorological conditions, historical fire occurrence patterns and plant species, and calibrate the correlation between each element;

[0009] S4: Based on the knowledge elements extracted in S3 and their associations, a knowledge graph of forest fire isolation zones is constructed, and the nodes and edges in the knowledge graph are semantically annotated to form a structured knowledge graph;

[0010] S5: Optimize the knowledge graph based on historical fire data and actual fire prevention experience, including adding new fire prevention strategies and removing redundant nodes and invalid edges.

[0011] Optionally, the S1 specifically includes:

[0012] S11: Obtain forest geographic distribution data through remote sensing satellites, use high-resolution image data to locate the boundaries, forest types and geographic distribution of forest areas, and generate corresponding forest geographic distribution layer data;

[0013] S12: Collect meteorological condition data, including temperature, humidity, wind speed and wind direction, through a meteorological sensor network and a meteorological data platform, and calibrate the collected meteorological condition data according to time and space coordinates;

[0014] S13: Collect historical fire data, including the time, location, fire spread speed and fire scale of the fire through the fire monitoring system and historical fire recording platform, and associate the collected historical fire data with the geographic information system;

[0015] S14: Obtain vegetation type distribution data by combining botanical surveys with remote sensing technology, classify and identify vegetation types in forest areas using remote sensing images, and calibrate the distribution locations of different plant species.

[0016] Optionally, the S2 specifically includes:

[0017] S21: De-noising the multi-source data collected in S11 to S14, using a high-pass filter algorithm to remove noise points in the image data, and using a Kalman filter algorithm to smooth the meteorological data and historical fire data to remove abnormal values ​​caused by environmental changes;

[0018] S22: Standardize the denoised data and use the Z-score standardization method to process the temperature, humidity, wind speed and wind direction indicators in the meteorological data and convert them into standardized data with zero mean and unit variance to eliminate the dimensional differences between different data sources;

[0019] S23: Fill missing values ​​in multi-source data using an interpolation method based on regression analysis. For missing values ​​in meteorological condition data, linear regression prediction is performed through known data and then filled. For missing values ​​in historical fire data, the nearest neighbor interpolation method is used to fill in the missing values ​​to ensure data integrity.

[0020] S24: Time alignment processing is performed on multi-source data. Various types of data are aligned through timestamps and a time window matching method is used to ensure the correspondence between meteorological data, fire data, and vegetation data on the same time axis;

[0021] S25: Integrate the multi-source data after denoising, standardization, missing value filling and time alignment into a unified format to generate the final forest fire isolation zone dataset.

[0022] Optionally, the S3 specifically includes:

[0023] S31: Extract geographic feature data from the forest fire isolation zone data set processed by S2, use geographic information system technology, based on geographic coordinates and terrain data, extract the boundaries of the forest area, terrain slope, forest type distribution and natural barriers, and mark the extracted geographic feature data as geographic feature nodes in the knowledge elements;

[0024] S32: Extract meteorological factors that affect fire spread from the forest fire isolation zone dataset, including temperature, humidity, wind speed and wind direction, and use principal component analysis to reduce the dimension of the data. Select the meteorological factor with the greatest impact on fire spread as the meteorological condition node in the knowledge element.

[0025] S33: Extract the spatial and temporal patterns of fire occurrence from the forest fire isolation zone dataset, use the cluster analysis algorithm to cluster the characteristics of historical fire frequency and fire spread pattern, identify the high-incidence areas and seasonal patterns of fire, and mark them as historical fire occurrence pattern nodes in the knowledge elements;

[0026] S34: Extract the species and distribution of various plants from the forest fire isolation belt dataset, use remote sensing image classification technology to refine the vegetation classification, extract flammable and fire-resistant plant species, and mark them as plant species nodes in the knowledge elements;

[0027] S35: Based on the various knowledge elements extracted from S31 to S34, an association rule mining algorithm is used to analyze and calibrate the associations between the various knowledge elements, including the association between geographical features and historical fire occurrence patterns, the association between meteorological conditions and fire spread speed, and the association between vegetation types and fire frequency.

[0028] Optionally, the S35 specifically includes:

[0029] S351: Based on the knowledge elements extracted from S31 to S34, an association rule mining algorithm is used to analyze the association between geographical features and the laws of historical fire occurrence. The specific steps include:

[0030] Calculate the support between geographical features and historical fire occurrence patterns. The support formula is: Among them, |X∩Y| represents the number of samples that contain both the geographic feature dataset X and the historical fire occurrence law dataset Y, and |D| represents the number of all samples;

[0031] Calculate the confidence between geographical features and historical fire occurrence patterns. The confidence formula is: Among them, Support(X) represents the frequency of occurrence of geographic feature X, represents the frequency of samples that contain both geographic feature X and historical fire occurrence pattern Y;

[0032] Calculate the lift between geographical features and historical fire occurrence patterns. The lift formula is: Among them, Support(X) represents the frequency of occurrence of geographical feature X, and Support(Y) represents the frequency of occurrence of historical fire occurrence pattern Y. Indicates the frequency of samples that contain both geographic feature X and historical fire occurrence pattern Y. Lift is used to reflect the strength of the association between geographic feature X and historical fire occurrence pattern Y. When Lift>1, it means that there is a positive correlation between X and Y.

[0033] S352: Analyze the correlation between meteorological conditions and fire spread speed using association rule mining algorithm. The specific steps include:

[0034] The support between meteorological conditions and fire spread speed is calculated as follows: Among them, |A∩B| represents the number of samples that contain both meteorological condition data A and fire spread speed data B, and |D| represents the number of all samples;

[0035] The confidence level between meteorological conditions and fire spread rate is calculated as follows: Among them, Support(A) represents the occurrence frequency of meteorological condition A, represents the frequency of samples that contain both meteorological condition A and fire spread speed B;

[0036] The formula for calculating the increase between meteorological conditions and fire spread speed is: Among them, Support(A) represents the occurrence frequency of meteorological condition A, Support(B) represents the occurrence frequency of fire spread speed B, Indicates the frequency of samples that contain both meteorological conditions A and fire spread speed B; Lift measures the strength of the association between meteorological conditions and fire spread speed. When Lift>1, it means that meteorological condition A has a positive impact on fire spread speed B;

[0037] S353: Using association rule mining algorithm to analyze the association between vegetation types and fire frequency, the specific steps include:

[0038] The support between vegetation types and fire frequency is calculated using the formula: Among them, |C∩D| represents the number of samples that contain both the vegetation type dataset C and the fire frequency dataset D, and |D| represents the number of all samples;

[0039] The confidence level between vegetation type and fire frequency is calculated using the formula: Among them, Support(C) represents the occurrence frequency of vegetation type C, represents the frequency of samples that contain both vegetation type c and fire frequency D;

[0040] The improvement between vegetation type and fire frequency is calculated using the formula: Among them, Support(C) represents the occurrence frequency of vegetation type C, Support(D) represents the occurrence frequency of fire frequency D, It represents the frequency of samples that contain both vegetation type C and fire frequency D. The lift measures the strength of the association between vegetation type and fire frequency. When Lift>1, it means that vegetation type C has a positive impact on fire frequency D.

[0041] Optionally, the S4 specifically includes:

[0042] S41: Determine the node set in the graph, where the nodes represent the key knowledge elements of the forest fire isolation zone, including geographical feature nodes, meteorological condition nodes, historical fire occurrence law nodes, and plant species nodes;

[0043] S42: Determine the edge set in the graph, where the edge represents the relationship between various knowledge elements, including the spatial and temporal relationship between geographical features and historical fire occurrence patterns, the relationship between meteorological conditions and fire spread speed, and the relationship between vegetation types and fire frequency;

[0044] S43: Using graph database technology, nodes and edges are stored according to the relationship between knowledge elements, and a forest fire isolation zone knowledge graph based on a graph model is constructed. The graph database uses a Neo4j storage model based on a graph structure to form a queryable and expandable knowledge graph by storing structured nodes and edges.

[0045] S44: Semantically annotate each node in the forest fire isolation belt knowledge graph, use natural language processing technology to parse the node description, and calibrate the specific meaning of the node; specifically, the semantic annotation of the geographical feature node is the forest geographical distribution, the semantic annotation of the meteorological condition node is the influence of meteorological factors on the spread of fire, the semantic annotation of the historical fire occurrence law node is the fire high-incidence area and seasonal law, and the semantic annotation of the plant species node is the plant species and its fire resistance;

[0046] S45: Calibrate the correlation between edges. Specifically, the correlation between geographical features and historical fire occurrence patterns is marked as the relationship between fire-prone areas and terrain slope, the correlation between meteorological conditions and fire spread speed is marked as the influence of wind speed on fire spread, and the correlation between plant species and fire frequency is marked as the relationship between flammable plants and fire frequency.

[0047] Optionally, the S43 specifically includes:

[0048] S431: Using Neo4j graph database technology, nodes and edges are stored according to the relationship between knowledge elements. Node entities in the graph database are first created. Node types include geographic feature nodes, meteorological condition nodes, historical fire occurrence law nodes, and plant species nodes. Each node contains attribute data related to the knowledge element.

[0049] S432: Create edge entities in the graph database, where the edges represent associations between nodes; each edge represents a relationship from one node to another node;

[0050] S433: Use the Cypher query language to manage and access nodes and edges through Neo4j's graph query engine, build a graph model based on the association relationship of knowledge elements, and perform data insertion, update and query operations;

[0051] S434: Store the nodes and edges in the Neo4j graph database, and build a query graph model to form a structured and queryable forest fire isolation zone knowledge graph.

[0052] Optionally, the S5 specifically includes:

[0053] S51: Based on historical fire data and fire spread rules, the association strength between nodes is evaluated. If the association strength between some nodes is lower than the preset threshold, the edge is considered invalid and marked and deleted. At the same time, the actual fire prevention experience is compared with the fire prevention measures data in the atlas, the redundant nodes in the atlas are identified, merged into one node, and the duplicate redundant nodes are deleted.

[0054] S52: Extract new fire prevention strategies from the latest historical fire data, including new isolation zone locations or fire prevention measures under different meteorological conditions; add the new fire prevention strategies as new knowledge element nodes to the knowledge graph, and associate them with the actual fire prevention experience and historical fire data to establish new edge relationships;

[0055] S53: After removing redundant nodes and invalid edges, the PageRank-based algorithm recalculates the weights of nodes and edges to ensure that all nodes and edges in the graph have practical significance and reference value.

[0056] Optionally, the S53 specifically includes:

[0057] S531: Initialize the initial weight of the node. The initial PageRank value of each node is set equal. The formula is: in, is the initial PageRank value of the ith node, and N2 is the total number of nodes in the graph;

[0058] S532: Construct a transition probability matrix between nodes to represent the structural characteristics of nodes and edges in the graph. The formula is: Among them, M ij is the probability of transferring from node j to node i, L(j) is the number of edges starting from node j;

[0059] S533: Iteratively calculate the PageRank value of each node, the formula is: in, is the PageRank value of the i-th node after the k+1th iteration, In(i) is the set of all nodes pointing to node i, L(j) is the number of outgoing edges of node j, and d is the damping coefficient (usually 0.85), which is used to control the impact of random jumps;

[0060] S534: Calculate the weight of the edge. The weight of the edge is defined based on the sum of the PageRank values ​​connecting the two nodes. The formula is: W ij =PR i +PR j , where W ij is the weight of the edge between node i and node j, PR i and pRj is the final PageRank value of the two nodes;

[0061] S535: Setting a convergence condition to check whether the change in the PageRank value is less than a preset threshold. If the condition is met, it is considered that the PageRank value has converged and the iteration ends.

[0062] S536: Sort the nodes and edges, prioritize high-weight nodes and edges, and delete low-weight redundant nodes and edges.

[0063] A forest fire isolation zone knowledge graph construction system, used to implement the above-mentioned forest fire isolation zone knowledge graph construction method, includes the following modules:

[0064] Data collection module: used to collect multi-source data related to forest fire prevention, including forest geographical distribution data, meteorological conditions data, historical fire data and vegetation type distribution data;

[0065] Data preprocessing module: connected to the data acquisition module, used to perform denoising, standardization, missing value filling and time alignment processing on the multi-source data collected by the data acquisition module to generate a forest fire isolation zone data set;

[0066] Knowledge extraction module: connected to the data preprocessing module, used to extract knowledge elements of geographical features, meteorological conditions, historical fire occurrence patterns and plant species from the forest fire isolation zone data set, and calibrate the correlation between each knowledge element;

[0067] Knowledge graph construction module: connected to the knowledge extraction module, constructs a forest fire isolation belt knowledge graph based on knowledge elements and their associations, and semantically annotates nodes and edges in the knowledge graph to form a structured knowledge graph;

[0068] Graph optimization module: connected to the knowledge graph construction module, used to optimize the knowledge graph according to historical fire data and actual fire prevention experience, including adding new fire prevention strategies and removing redundant nodes and invalid edges.

[0069] Beneficial effects of the present invention:

[0070] The present invention, through multi-source data fusion technology, can accurately integrate various information such as forest geographical distribution, meteorological conditions, historical fire data, and vegetation types, so as to effectively construct a comprehensive forest fire isolation zone knowledge map; through this knowledge map, the various factors affecting the spread of fire and their interrelationships can be comprehensively analyzed, providing a scientific basis for the planning and optimization of fire isolation zones.

[0071] The present invention optimizes the nodes and edges in the knowledge graph by introducing the PageRank algorithm, thereby ensuring the efficiency and scalability of the graph structure; with the support of historical fire data, it can continuously update and optimize fire prevention strategies, remove redundant data and invalid associations, thereby maintaining the efficiency and real-time performance of the knowledge graph; this optimization method not only improves the adaptability of the graph under different fire environments, but also enhances the scientific nature and operability of fire prevention decisions, providing more accurate technical support for forest fire prevention work. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0073] Figure 1 A schematic diagram of a method for constructing a knowledge graph of a forest fire isolation zone according to an embodiment of the present invention;

[0074] Figure 2 A schematic diagram of a system for constructing a knowledge graph of forest fire isolation zones according to an embodiment of the present invention. DETAILED DESCRIPTION

[0075] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0076] It should be noted that the references to "one embodiment", "an embodiment", "an exemplary embodiment", "some embodiments" and the like in the specification indicate that the embodiments described may include specific features, structures or characteristics, but not every embodiment may include the specific features, structures or characteristics. In addition, when a specific feature, structure or characteristic is described in conjunction with an embodiment, it should be within the knowledge of a person skilled in the art to implement such feature, structure or characteristic in conjunction with other embodiments (whether or not explicitly described).

[0077] In general, a term can be understood, at least in part, from its use in context. For example, depending, at least in part, on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending, at least in part, on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0078] like Figure 1 As shown, a method for constructing a knowledge graph of a forest fire isolation zone includes the following steps:

[0079] S1: Collect multi-source data related to forest fire prevention, including forest geographical distribution data, meteorological conditions data, historical fire data and vegetation type distribution data;

[0080] S2: De-noise, standardize, fill in missing values ​​and align time for the multi-source data collected in S1 to generate a forest fire isolation zone dataset;

[0081] S3: Extract knowledge elements about forest fire isolation belts from the forest fire isolation belt dataset, including geographical features, meteorological conditions, historical fire occurrence patterns and plant species, and calibrate the correlation between each element;

[0082] S4: Based on the knowledge elements extracted in S3 and their associations, a knowledge graph of forest fire isolation zones is constructed, and the nodes and edges in the knowledge graph are semantically annotated to form a structured knowledge graph;

[0083] S5: Based on historical fire data and actual fire prevention experience, the knowledge graph is optimized, including adding new fire prevention strategies and removing redundant nodes and invalid edges to ensure the accuracy and effectiveness of the graph, and to obtain a forest fire isolation zone knowledge graph model that meets actual fire prevention needs.

[0084] S1 specifically includes:

[0085] S11: Obtain forest geographic distribution data through remote sensing satellites, use high-resolution image data to locate the boundaries, forest types and geographic distribution of forest areas, and generate corresponding forest geographic distribution layer data;

[0086] S12: Collect meteorological condition data, including temperature, humidity, wind speed and wind direction, through a meteorological sensor network and a meteorological data platform, and calibrate the collected meteorological condition data according to time and space coordinates;

[0087] S13: Collect historical fire data, including the time, location, fire spread speed and fire scale of the fire through the fire monitoring system and historical fire recording platform, and associate the collected historical fire data with the geographic information system;

[0088] S14: Through the combination of botanical surveys and remote sensing technology, vegetation type distribution data is obtained, remote sensing images are used to classify and identify vegetation types in forest areas, and the distribution positions of different plant species are calibrated; through the implementation of the above technical steps, various types of forest fire prevention related data can be obtained efficiently and accurately, including forest geographical distribution, meteorological conditions, historical fires, and vegetation types, etc., ensuring the comprehensiveness and accuracy of the data, and providing a solid data foundation for subsequent data preprocessing and knowledge graph construction.

[0089] S2 specifically includes:

[0090] S21: De-noising the multi-source data collected in S11 to S14, using a high-pass filter algorithm to remove noise points in the image data, and using a Kalman filter algorithm to smooth the meteorological data and historical fire data to remove abnormal values ​​caused by environmental changes;

[0091] S22: Standardize the denoised data and use the Z-score standardization method to process the temperature, humidity, wind speed and wind direction indicators in the meteorological data and convert them into standardized data with zero mean and unit variance to eliminate the dimensional differences between different data sources;

[0092] S23: Fill missing values ​​in multi-source data using an interpolation method based on regression analysis. For missing values ​​in meteorological condition data, linear regression prediction is performed through known data and then filled. For missing values ​​in historical fire data, the nearest neighbor interpolation method (KNN interpolation) is used to fill in the missing values ​​to ensure data integrity.

[0093] S24: Time alignment of multi-source data is performed, and various types of data are aligned through timestamps. The time window matching method is used to ensure the correspondence between meteorological data, fire data, and vegetation data on the same time axis, thereby eliminating the time deviation caused by different data collection frequencies;

[0094] S25: Integrate the multi-source data after denoising, standardization, missing value filling and time alignment into a unified format to generate the final forest fire isolation zone data set, providing processed and well-input data for subsequent knowledge extraction and graph construction; through the above technical steps, it is possible to effectively process a variety of data collected from different sources to ensure the accuracy, consistency and completeness of the data; after denoising, standardization, missing value filling and time alignment, the data will be more suitable for subsequent knowledge extraction and graph construction, improving the construction quality and accuracy of the forest fire isolation zone knowledge graph.

[0095] S3 specifically includes:

[0096] S31: Extract geographic feature data from the forest fire isolation zone data set processed by S2, use geographic information system (GIS) technology, based on geographic coordinates and terrain data, extract the boundaries of the forest area, terrain slope, forest type distribution and natural barriers (such as rivers, mountains, etc.), and mark the extracted geographic feature data as geographic feature nodes in the knowledge elements; the extraction of the above geographic feature data can be calculated by the following formula: Among them, ΔZ is the height difference between adjacent geographical points, and ΔX is the horizontal distance; the terrain slope can reflect the changes in the terrain height of the forest area, and thus provide a basis for the site selection of the fire isolation zone;

[0097] S32: Extract meteorological factors that affect fire spread from the forest fire isolation zone data set, including temperature, humidity, wind speed and wind direction, and use the principal component analysis (PCA) method to reduce the dimension of the data, and select the meteorological factors that have the greatest impact on fire spread as the meteorological condition nodes in the knowledge elements; the core formula of the PCA dimensionality reduction process is: Z = X·W, where X is the original data matrix, W is the eigenvector matrix, and Z is the data after dimensionality reduction; through this formula, the main influencing factors can be extracted, and redundant information can be reduced, so as to focus on the meteorological conditions that have the greatest impact on fire spread;

[0098] S33: Extract the spatial and temporal patterns of fire occurrence from the forest fire isolation zone dataset, use the clustering analysis algorithm to cluster the characteristics of historical fire frequency and fire spread pattern, identify the high-incidence areas and seasonal patterns of fire, and mark them as historical fire occurrence pattern nodes in the knowledge elements; the K-means algorithm used by the clustering algorithm is performed through the following formula: Among them, J is the clustering cost function, x i is the i-th sample point, μ k is the kth cluster center, and K is the number of clusters. Through this algorithm, high-incidence areas of fire and seasonal patterns can be identified, providing a reference for the layout of fire isolation zones.

[0099] S34: Extract the species and distribution of various plants from the forest fire isolation belt dataset, use remote sensing image classification technology to refine the vegetation classification, extract flammable and fire-resistant plant species, and mark them as plant species nodes in the knowledge elements; the formula used for vegetation classification in remote sensing image classification is: Among them, NDVI is the Normalized Difference Vegetation Index, NIR is the near-infrared band reflectance, and RED is the red light band reflectance. The NDVI value can effectively distinguish different types of vegetation and calibrate flammable plants and fire-resistant plants.

[0100] S35: Based on the various knowledge elements extracted from S31 to S34, an association rule mining algorithm is used to analyze and calibrate the associations between the various knowledge elements, including the association between geographical features and historical fire occurrence patterns, the association between meteorological conditions and fire spread speed, and the association between vegetation types and fire occurrence frequency, to form meaningful relationships between elements, and represent them as edge nodes in the graph, thereby completing the calibration of the associations between knowledge elements; through the above technical steps, the key knowledge elements that affect forest fire prevention can be systematically and accurately extracted from multi-source data, including geographical features, meteorological conditions, historical fire patterns and vegetation types; and through the association rule mining algorithm, the intrinsic relationships between these knowledge elements can be effectively identified and calibrated, providing a scientific basis for the subsequent construction of a knowledge graph for forest fire isolation zones.

[0101] S35 specifically includes:

[0102] S351: Based on the knowledge elements extracted from S31 to S34, an association rule mining algorithm is used to analyze the association between geographical features and the laws of historical fire occurrence. The specific steps include:

[0103] Calculate the support between geographical features and historical fire occurrence patterns. The support formula is: Among them, |X∩Y| represents the number of samples that contain both the geographic feature dataset X and the historical fire occurrence law dataset Y, and |D| represents the number of all samples;

[0104] Calculate the confidence between geographical features and historical fire occurrence patterns. The confidence formula is: Among them, Support(X) represents the frequency of occurrence of geographic feature X, represents the frequency of samples that contain both geographic feature X and historical fire occurrence pattern Y;

[0105] Calculate the lift between geographical features and historical fire occurrence patterns. The lift formula is: Among them, Support(X) represents the frequency of occurrence of geographical feature X, and Support(Y) represents the frequency of occurrence of historical fire occurrence pattern Y. Indicates the frequency of samples that contain both geographic feature X and historical fire occurrence pattern Y. Lift is used to reflect the strength of the association between geographic feature X and historical fire occurrence pattern Y. When Lift>1, it means that there is a positive correlation between X and Y.

[0106] S352: Analyze the correlation between meteorological conditions and fire spread speed using association rule mining algorithm. The specific steps include:

[0107] Calculate the support between meteorological conditions and fire spread speed (Support), the formula is: Among them, |A∩B| represents the number of samples that contain both meteorological condition data A and fire spread speed data B, and |D| represents the number of all samples;

[0108] The confidence between meteorological conditions and fire spread speed is calculated using the formula: Among them, Support(A) represents the occurrence frequency of meteorological condition A, represents the frequency of samples that contain both meteorological condition A and fire spread speed B;

[0109] The lift between meteorological conditions and fire spread speed is calculated using the formula: Among them, Support(A) represents the occurrence frequency of meteorological condition A, Support(B) represents the occurrence frequency of fire spread speed B, Indicates the frequency of samples that contain both meteorological conditions A and fire spread speed B; Lift measures the strength of the association between meteorological conditions and fire spread speed. When Lift>1, it means that meteorological condition A has a positive impact on fire spread speed B;

[0110] S353: Using association rule mining algorithm to analyze the association between vegetation types and fire frequency, the specific steps include:

[0111] The support between vegetation types and fire frequency is calculated using the formula: Among them, |C∩D| represents the number of samples that contain both the vegetation type dataset C and the fire frequency dataset D, and |D| represents the number of all samples;

[0112] The confidence between vegetation type and fire frequency is calculated using the formula: Among them, Support(C) represents the occurrence frequency of vegetation type C, represents the frequency of samples containing both vegetation type C and fire frequency D;

[0113] The lift between vegetation type and fire frequency is calculated using the formula: Among them, Support(C) represents the occurrence frequency of vegetation type C, Support(D) represents the occurrence frequency of fire frequency D, represents the frequency of samples that contain both vegetation type C and fire frequency D; lift measures the strength of the association between vegetation type and fire frequency. When Lift>1, it means that vegetation type C has a positive impact on fire frequency D. Through the above association rule mining algorithm, the intrinsic connection between different knowledge elements can be efficiently identified, and how factors such as geographical characteristics, meteorological conditions, historical fire laws and vegetation types interact with each other to affect the effectiveness of forest fire isolation belts can be further revealed. These associations provide a scientific basis for the optimal design of fire isolation belts and the formulation of fire prevention strategies.

[0114] S4 specifically includes:

[0115] S41: Determine the node set in the graph, where the nodes represent the key knowledge elements of the forest fire isolation zone, including geographical feature nodes, meteorological condition nodes, historical fire occurrence law nodes, and plant species nodes;

[0116] S42: Determine the edge set in the graph, where the edge represents the relationship between various knowledge elements, including the spatial and temporal relationship between geographical features and historical fire occurrence patterns, the relationship between meteorological conditions and fire spread speed, and the relationship between vegetation types and fire frequency;

[0117] S43: Using graph database technology, nodes and edges are stored according to the relationship between knowledge elements, and a knowledge graph of forest fire isolation zones based on a graph model is constructed. The graph database uses a Neo4j storage model based on a graph structure to form a queryable and extensible knowledge graph by storing structured nodes and edges.

[0118] S44: Semantically annotate each node in the forest fire isolation belt knowledge graph, use natural language processing (NLP) technology to parse the node description, and calibrate the specific meaning of the node; specifically, the semantic annotation of the geographical feature node is the forest geographical distribution, the semantic annotation of the meteorological condition node is the impact of meteorological factors on the spread of fire, the semantic annotation of the historical fire occurrence law node is the fire high-incidence area and seasonal law, and the semantic annotation of the plant species node is the plant species and its fire resistance;

[0119] S45: Calibrate the correlation relationships of the edges. Specifically, the correlation between geographical features and historical fire occurrence patterns is marked as the relationship between fire-prone areas and terrain slope, the correlation between meteorological conditions and fire spread speed is marked as the impact of wind speed on fire spread, and the correlation between plant species and fire frequency is marked as the relationship between flammable plants and fire frequency. Through the above steps, a forest fire isolation zone knowledge graph model based on multiple data sources and correlation relationships can be effectively constructed, and the nodes and edges in the graph can be accurately semantically labeled. This structured knowledge graph can not only accurately express various key factors and their relationships in the field of forest fire prevention, but also provide a precise knowledge basis for subsequent fire prevention strategy optimization and decision support.

[0120] S43 specifically includes:

[0121] S431: Using Neo4j graph database technology, nodes and edges are stored according to the relationship between knowledge elements. First, node entities in the graph database are created. Node types include geographic feature nodes, meteorological condition nodes, historical fire occurrence law nodes, and plant species nodes. Each node contains attribute data related to the knowledge element, expressed as: node attribute = {node ID, node name, associated data}, where the node ID is a unique identifier, the node name indicates the node type (such as geographic features, meteorological conditions, etc.), and the associated data includes detailed information related to the corresponding node (such as coordinates, meteorological data, fire occurrence frequency, etc.);

[0122] S432: Create edge entities in the graph database. Edges represent associations between nodes, such as the relationship between geographical features and historical fire occurrence patterns, the relationship between meteorological conditions and fire spread speed, and the relationship between vegetation types and fire frequency. Each edge represents a relationship from one node to another. The following formula is used to define the attributes of the edge: edge attribute = {edge ID, starting node ID, target node ID, relationship type, association strength}, where the starting node ID and the target node ID are the IDs of the associated nodes, respectively, the relationship type is the type of edge (such as "spatial association", "temporal association", "influence relationship", etc.), and the association strength represents the strength of the relationship between nodes. For example, the association strength between fire-prone areas and terrain slope can be obtained by analyzing historical fire data.

[0123] S433: Use the Cypher query language to manage and access nodes and edges through Neo4j's graph query engine, build a graph model based on the association relationship of knowledge elements, and perform data insertion, update and query operations;

[0124] S434: Store the nodes and edges in the Neo4j graph database, and build a query graph model to form a structured and queryable forest fire isolation zone knowledge graph to facilitate subsequent analysis and optimization.

[0125] The optimization of knowledge graph in S5 specifically includes:

[0126] S51: Based on historical fire data and actual fire prevention experience, analyze the effectiveness of each node and edge in the current forest fire isolation belt knowledge graph, identify redundant nodes and invalid edges in the graph through data mining technology, and evaluate the association strength between nodes based on historical fire data and fire spread laws. If the association strength between some nodes is lower than the preset threshold, the edge is considered invalid and marked and deleted. At the same time, compare the actual fire prevention experience with the fire prevention measures data in the graph to identify redundant nodes in the graph. If multiple nodes represent similar geographical features or meteorological conditions, merge them into one node and delete duplicate redundant nodes.

[0127] S52: Extract new fire prevention strategies from the latest historical fire data, including new isolation zone locations or fire prevention measures under different meteorological conditions; add the new fire prevention strategies as new knowledge element nodes to the knowledge graph, and associate them with the historical fire data based on actual fire prevention experience to establish new edge relationships; the creation of new nodes is represented by: new node attributes = {node ID, fire prevention strategy name, applicable area, validity period, associated data}, where the node ID is the unique identifier of the new node, the fire prevention strategy name indicates the new fire prevention strategy content, the applicable area is the geographical area to which the strategy applies, and the associated data includes detailed information related to the strategy, such as implementation effects, etc.;

[0128] S53: After removing redundant nodes and invalid edges, the PageRank-based algorithm recalculates the weights of nodes and edges to ensure that all nodes and edges in the graph have practical significance and reference value. The optimized graph can more accurately reflect the effectiveness of forest fire isolation zones.

[0129] S53 specifically includes:

[0130] S531: Initialize the initial weight of the node. The initial PageRank value of each node is set equal. The formula is: in, is the initial PageRank value of the ith node, and N2 is the total number of nodes in the graph;

[0131] S532: Construct a transition probability matrix between nodes to represent the structural characteristics of nodes and edges in the graph. The formula is: Among them, M ij is the probability of transferring from node j to node i, L(j) is the number of edges starting from node j;

[0132] S533: Iteratively calculate the PageRank value of each node, the formula is: in, is the PageRank value of the i-th node after the k+1th iteration, In(i) is the set of all nodes pointing to node i, L(j) is the number of outgoing edges of node j, and d is the damping coefficient (usually 0.85), which is used to control the impact of random jumps;

[0133] S534: Calculate the weight of the edge. The weight of the edge is defined based on the sum of the PageRank values ​​connecting the two nodes. The formula is: W ij =PR i +PR j , where W ij is the weight of the edge between node i and node j, PR i and pR j is the final PageRank value of the two nodes;

[0134] S535: Setting a convergence condition to check whether the change in the PageRank value is less than a preset threshold. If the condition is met, it is considered that the PageRank value has converged and the iteration ends.

[0135] S536: Sort the nodes and edges, give priority to retaining high-weight nodes and edges, and delete redundant nodes and edges with low weights; the above steps can quantify the importance of each node and edge in the knowledge graph by adopting the PageRank algorithm, ensuring that the optimized graph can reflect the actual structure and key correlation relationships of the forest fire isolation zone. This method provides a scientific weight evaluation standard, which helps to delete redundant information, highlight key knowledge elements, and improve the accuracy and practicality of the knowledge graph.

[0136] like Figure 2 As shown, a forest fire isolation zone knowledge graph construction system is used to implement the above-mentioned forest fire isolation zone knowledge graph construction method, including the following modules:

[0137] Data collection module: used to collect multi-source data related to forest fire prevention, including forest geographical distribution data, meteorological conditions data, historical fire data and vegetation type distribution data;

[0138] Data preprocessing module: connected to the data acquisition module, used to perform denoising, standardization, missing value filling and time alignment on the multi-source data collected by the data acquisition module to generate a forest fire isolation zone data set;

[0139] Knowledge extraction module: connected to the data preprocessing module, used to extract knowledge elements of geographical features, meteorological conditions, historical fire occurrence patterns and plant species from the forest fire isolation zone data set, and calibrate the correlation between each knowledge element;

[0140] Knowledge graph construction module: connected to the knowledge extraction module, constructs a forest fire isolation belt knowledge graph based on knowledge elements and their associations, and semantically annotates nodes and edges in the knowledge graph to form a structured knowledge graph;

[0141] Graph optimization module: connected to the knowledge graph construction module, used to optimize the knowledge graph based on historical fire data and actual fire prevention experience, including adding new fire prevention strategies and removing redundant nodes and invalid edges to ensure the accuracy and effectiveness of the knowledge graph; through the collaborative work of the various modules of the above system, it is possible to achieve the collection and preprocessing of multi-source data, the extraction and association of knowledge elements, the construction and semantic annotation of the knowledge graph, and the optimization of the knowledge graph, forming a complete, dynamic and structured forest fire isolation zone knowledge graph system.

[0142] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0143] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for constructing a knowledge graph of forest fire isolation zones, characterized in that: The following steps are involved: S1: Collect multi-source data related to forest fire prevention, including forest geographical distribution data, meteorological conditions data, historical fire data and vegetation type distribution data; S2: De-noise, standardize, fill in missing values ​​and align time for the multi-source data collected in S1 to generate a forest fire isolation zone dataset; S3: Extract knowledge elements about forest fire isolation belts from the forest fire isolation belt dataset, including geographical features, meteorological conditions, historical fire occurrence patterns and plant species, and calibrate the correlation between each element; S4: Based on the knowledge elements extracted in S3 and their associations, a knowledge graph of forest fire isolation zones is constructed, and the nodes and edges in the knowledge graph are semantically annotated to form a structured knowledge graph; S5: Optimize the knowledge graph based on historical fire data and actual fire prevention experience, including adding new fire prevention strategies and removing redundant nodes and invalid edges.

2. A method for constructing a knowledge graph of forest fire isolation zones according to claim 1, characterized in that: The S1 specifically includes: S11: Obtain forest geographic distribution data through remote sensing satellites, use high-resolution image data to locate the boundaries, forest types and geographic distribution of forest areas, and generate corresponding forest geographic distribution layer data; S12: Collect meteorological condition data, including temperature, humidity, wind speed and wind direction, through a meteorological sensor network and a meteorological data platform, and calibrate the collected meteorological condition data according to time and space coordinates; S13: Collect historical fire data, including the time, location, fire spread speed and fire scale of the fire through the fire monitoring system and historical fire recording platform, and associate the collected historical fire data with the geographic information system; S14: Obtain vegetation type distribution data by combining botanical surveys with remote sensing technology, classify and identify vegetation types in forest areas using remote sensing images, and calibrate the distribution locations of different plant species.

3. A method for constructing a knowledge graph of forest fire isolation zones according to claim 1, characterized in that: The S2 specifically includes: S21: De-noising the multi-source data collected in S11 to S14, using a high-pass filter algorithm to remove noise points in the image data, and using a Kalman filter algorithm to smooth the meteorological data and historical fire data to remove abnormal values ​​caused by environmental changes; S22: Standardize the denoised data and use the Z-score standardization method to process the temperature, humidity, wind speed and wind direction indicators in the meteorological data and convert them into standardized data with zero mean and unit variance to eliminate the dimensional differences between different data sources; S23: Fill missing values ​​in multi-source data using an interpolation method based on regression analysis. For missing values ​​in meteorological condition data, linear regression prediction is performed through known data and then filled. For missing values ​​in historical fire data, the nearest neighbor interpolation method is used to fill in the missing values ​​to ensure data integrity. S24: Time alignment processing is performed on multi-source data. Various types of data are aligned through timestamps and a time window matching method is used to ensure the correspondence between meteorological data, fire data, and vegetation data on the same time axis; S25: Integrate the multi-source data after denoising, standardization, missing value filling and time alignment into a unified format to generate the final forest fire isolation zone dataset.

4. A method for constructing a knowledge graph of forest fire isolation zones according to claim 1, characterized in that: The S3 specifically includes: S31: Extract geographic feature data from the forest fire isolation zone data set processed by S2, use geographic information system technology, based on geographic coordinates and terrain data, extract the boundaries of the forest area, terrain slope, forest type distribution and natural barriers, and mark the extracted geographic feature data as geographic feature nodes in the knowledge elements; S32: Extract meteorological factors that affect fire spread from the forest fire isolation zone dataset, including temperature, humidity, wind speed and wind direction, and use principal component analysis to reduce the dimension of the data. Select the meteorological factor with the greatest impact on fire spread as the meteorological condition node in the knowledge element. S33: Extract the spatial and temporal patterns of fire occurrence from the forest fire isolation zone dataset, use the cluster analysis algorithm to cluster the characteristics of historical fire frequency and fire spread pattern, identify the high-incidence areas and seasonal patterns of fire, and mark them as historical fire occurrence pattern nodes in the knowledge elements; S34: Extract the species and distribution of various plants from the forest fire isolation belt dataset, use remote sensing image classification technology to refine the vegetation classification, extract flammable and fire-resistant plant species, and mark them as plant species nodes in the knowledge elements; S35: Based on the various knowledge elements extracted from S31 to S34, an association rule mining algorithm is used to analyze and calibrate the associations between the various knowledge elements, including the association between geographical features and historical fire occurrence patterns, the association between meteorological conditions and fire spread speed, and the association between vegetation types and fire frequency.

5. A method for constructing a knowledge graph of forest fire isolation zones according to claim 4, characterized in that: The S35 specifically includes: S351: Based on the knowledge elements extracted from S31 to S34, an association rule mining algorithm is used to analyze the association between geographical features and the laws of historical fire occurrence. The specific steps include: Calculate the support between geographical features and historical fire occurrence patterns. The support formula is: Among them, |X∩Y| represents the number of samples that contain both the geographic feature dataset X and the historical fire occurrence law dataset Y, and |D| represents the number of all samples; Calculate the confidence between geographical features and historical fire occurrence patterns. The confidence formula is: Among them, Support(X) represents the frequency of occurrence of geographic feature X, represents the frequency of samples that contain both geographic feature X and historical fire occurrence pattern Y; Calculate the lift between geographical features and historical fire occurrence patterns. The lift formula is: Among them, Support(X) represents the frequency of occurrence of geographical feature X, and Support(Y) represents the frequency of occurrence of historical fire occurrence pattern Y. Indicates the frequency of samples that contain both geographic feature X and historical fire occurrence pattern Y. Lift is used to reflect the strength of the association between geographic feature X and historical fire occurrence pattern Y. When Lift>1, it means that there is a positive correlation between X and Y. S352: Analyze the correlation between meteorological conditions and fire spread speed using association rule mining algorithm. The specific steps include: The support between meteorological conditions and fire spread speed is calculated as follows: Among them, |A∩B| represents the number of samples that contain both meteorological condition data A and fire spread speed data B, and |D| represents the number of all samples; The confidence level between meteorological conditions and fire spread rate is calculated as follows: Among them, Support(A) represents the occurrence frequency of meteorological condition A, represents the frequency of samples that contain both meteorological condition A and fire spread speed B; The formula for calculating the increase between meteorological conditions and fire spread speed is: Among them, Support(A) represents the occurrence frequency of meteorological condition A, Support(B) represents the occurrence frequency of fire spread speed B, Indicates the frequency of samples that contain both meteorological conditions A and fire spread speed B; Lift measures the strength of the association between meteorological conditions and fire spread speed. When Lift>1, it means that meteorological condition A has a positive impact on fire spread speed B; S353: Using association rule mining algorithm to analyze the association between vegetation types and fire frequency, the specific steps include: The support between vegetation types and fire frequency is calculated using the formula: Among them, |C∩D| represents the number of samples that contain both the vegetation type dataset C and the fire frequency dataset D, and |D| represents the number of all samples; The confidence level between vegetation type and fire frequency is calculated using the formula: Among them, Support(C) represents the occurrence frequency of vegetation type C, represents the frequency of samples containing both vegetation type C and fire frequency D; The improvement between vegetation type and fire frequency is calculated using the formula: Among them, Support(C) represents the occurrence frequency of vegetation type C, Support(D) represents the occurrence frequency of fire frequency D, It represents the frequency of samples that contain both vegetation type C and fire frequency D. The lift measures the strength of the association between vegetation type and fire frequency. When Lift>1, it means that vegetation type C has a positive impact on fire frequency D.

6. A method for constructing a knowledge graph of forest fire isolation zones according to claim 1, characterized in that: The S4 specifically includes: S41: Determine the node set in the graph, where the nodes represent the key knowledge elements of the forest fire isolation zone, including geographical feature nodes, meteorological condition nodes, historical fire occurrence law nodes, and plant species nodes; S42: Determine the edge set in the graph, where the edge represents the relationship between various knowledge elements, including the spatial and temporal relationship between geographical features and historical fire occurrence patterns, the relationship between meteorological conditions and fire spread speed, and the relationship between vegetation types and fire frequency; S43: Using graph database technology, nodes and edges are stored according to the relationship between knowledge elements, and a forest fire isolation zone knowledge graph based on a graph model is constructed. The graph database uses a Neo4j storage model based on a graph structure to form a queryable and expandable knowledge graph by storing structured nodes and edges. S44: Semantically annotate each node in the forest fire isolation belt knowledge graph, use natural language processing technology to parse the node description, and calibrate the specific meaning of the node; specifically, the semantic annotation of the geographical feature node is the forest geographical distribution, the semantic annotation of the meteorological condition node is the influence of meteorological factors on the spread of fire, the semantic annotation of the historical fire occurrence law node is the fire high-incidence area and seasonal law, and the semantic annotation of the plant species node is the plant species and its fire resistance; S45: Calibrate the correlation between edges. Specifically, the correlation between geographical features and historical fire occurrence patterns is marked as the relationship between fire-prone areas and terrain slope, the correlation between meteorological conditions and fire spread speed is marked as the influence of wind speed on fire spread, and the correlation between plant species and fire frequency is marked as the relationship between flammable plants and fire frequency.

7. A method for constructing a knowledge graph of forest fire isolation zones according to claim 6, characterized in that: The S43 specifically includes: S431: Using Neo4j graph database technology, nodes and edges are stored according to the relationship between knowledge elements. Node entities in the graph database are first created. Node types include geographic feature nodes, meteorological condition nodes, historical fire occurrence law nodes, and plant species nodes. Each node contains attribute data related to the knowledge element. S432: Create edge entities in the graph database, where the edges represent associations between nodes; each edge represents a relationship from one node to another node; S433: Use the Cypher query language to manage and access nodes and edges through Neo4j's graph query engine, build a graph model based on the association relationship of knowledge elements, and perform data insertion, update and query operations; S434: Store the nodes and edges in the Neo4j graph database, and build a query graph model to form a structured and queryable forest fire isolation zone knowledge graph.

8. The method for constructing a knowledge graph of a forest fire isolation zone according to claim 1, characterized in that: The S5 specifically includes: S51: Based on historical fire data and fire spread rules, the association strength between nodes is evaluated. If the association strength between some nodes is lower than the preset threshold, the edge is considered invalid and marked and deleted. At the same time, the actual fire prevention experience is compared with the fire prevention measures data in the atlas, the redundant nodes in the atlas are identified, merged into one node, and the duplicate redundant nodes are deleted. S52: Extract new fire prevention strategies from the latest historical fire data, including new isolation zone locations or fire prevention measures under different meteorological conditions; add the new fire prevention strategies as new knowledge element nodes to the knowledge graph, and associate them with the actual fire prevention experience and historical fire data to establish new edge relationships; S53: After removing redundant nodes and invalid edges, the PageRank-based algorithm recalculates the weights of nodes and edges to ensure that all nodes and edges in the graph have practical significance and reference value.

9. A method for constructing a knowledge graph of forest fire isolation zones according to claim 8, characterized in that: The S53 specifically includes: S531: Initialize the initial weight of the node. The initial PageRank value of each node is set equal. The formula is: in, is the initial PageRank value of the ith node, and N2 is the total number of nodes in the graph; S532: Construct a transition probability matrix between nodes to represent the structural characteristics of nodes and edges in the graph. The formula is: Among them, M ij is the probability of transferring from node j to node i, L(j) is the number of edges starting from node j; S533: Iteratively calculate the PageRank value of each node, the formula is: in, is the PageRank value of the i-th node after the k+1th iteration, In(i) is the set of all nodes pointing to node i, L(j) is the number of outgoing edges of node j, and d is the damping coefficient (usually 0.85), which is used to control the impact of random jumps; S534: Calculate the weight of the edge. The weight of the edge is defined based on the sum of the PageRank values ​​connecting the two nodes. The formula is: W ij =PR i +PR j , where W ij is the weight of the edge between node i and node j, PR i and PR j is the final PageRank value of the two nodes; S535: Setting a convergence condition to check whether the change in the PageRank value is less than a preset threshold. If the condition is met, it is considered that the PageRank value has converged and the iteration ends. S536: Sort the nodes and edges, prioritize high-weight nodes and edges, and delete low-weight redundant nodes and edges.

10. A forest fire isolation zone knowledge graph construction system, used to implement a forest fire isolation zone knowledge graph construction method as claimed in any one of claims 1 to 9, characterized in that: Includes the following modules: Data collection module: used to collect multi-source data related to forest fire prevention, including forest geographical distribution data, meteorological conditions data, historical fire data and vegetation type distribution data; Data preprocessing module: connected to the data acquisition module, used to perform denoising, standardization, missing value filling and time alignment processing on the multi-source data collected by the data acquisition module to generate a forest fire isolation zone data set; Knowledge extraction module: connected to the data preprocessing module, used to extract knowledge elements of geographical features, meteorological conditions, historical fire occurrence patterns and plant species from the forest fire isolation zone data set, and calibrate the correlation between each knowledge element; Knowledge graph construction module: connected to the knowledge extraction module, constructs a forest fire isolation belt knowledge graph based on knowledge elements and their associations, and semantically annotates nodes and edges in the knowledge graph to form a structured knowledge graph; Graph optimization module: connected to the knowledge graph construction module, used to optimize the knowledge graph according to historical fire data and actual fire prevention experience, including adding new fire prevention strategies and removing redundant nodes and invalid edges.

Images