Meteorological risk early warning method and system applied to geological disaster analysis

Abstract

The embodiment of the application discloses a meteorological risk early warning method and system applied to geological disaster analysis, the method comprises the following steps: performing numerical meteorological change analysis processing on meteorological element monitoring data of a target geological area to generate meteorological change analysis results; then, combining a GIS spatial analysis strategy, performing fusion early warning processing on the meteorological change analysis results and geological radar detection information of the target geological area to obtain geological disaster risk early warning results; then, establishing a meteorological geological risk early warning knowledge base based on the meteorological element monitoring data, the geological radar detection information and the geological disaster risk early warning results; finally, performing geological disaster risk assessment processing on real-time meteorological monitoring data of the target geological area by using the knowledge base to output real-time geological disaster risk early warning information, and the method realizes deep fusion of meteorological and geological information, and improves the accuracy and real-time performance of geological disaster early warning.

Description

Technical Field

[0001] This invention relates to the field of geological disaster analysis technology, specifically to a meteorological risk early warning method and system for geological disaster analysis. Background Technology

[0002] In geological hazard analysis, existing technologies have been used in various research and practice, primarily assessing the risk of geological hazards by analyzing single meteorological elements or geological conditions. Meteorologically, the focus is on changes in meteorological elements such as rainfall and temperature, using established thresholds to determine the likelihood of geological hazards. When rainfall exceeds a preset value, it is considered likely to trigger geological hazards such as landslides and debris flows. Geologically, geological survey reports are mainly relied upon to understand the basic geological structure and to conduct a preliminary assessment of the potential risks of geological hazards.

[0003] However, most existing methods analyze meteorological and geological information independently or perform simple correlations, lacking in-depth exploration and comprehensive consideration of the complex interactions between the two. The impact of changes in meteorological elements on geological structures is a dynamic and complex process, and existing methods struggle to accurately capture this dynamic change, resulting in low accuracy and reliability of early warnings and making it difficult to conduct rapid and accurate geological hazard risk assessments. Summary of the Invention

[0004] This invention provides a meteorological risk early warning method and system for geological disaster analysis.

[0005] In a first aspect, embodiments of the present invention provide a meteorological risk early warning method for geological disaster analysis, the meteorological risk early warning method for geological disaster analysis comprising:

[0006] Numerical meteorological change analysis is performed on meteorological element monitoring data of the target geological area to generate meteorological change analysis results.

[0007] By combining the results of meteorological change analysis with the geological radar detection information of the target geological area using GIS spatial analysis strategies, a geological disaster risk warning result for the target geological area is obtained.

[0008] Establish a meteorological and geological risk early warning knowledge base based on meteorological element monitoring data, geological radar detection information, and geological disaster risk early warning results;

[0009] The meteorological and geological risk early warning knowledge base is used to conduct geological disaster risk assessment on the real-time meteorological monitoring data of the target geological area and output real-time geological disaster risk early warning information for the target geological area.

[0010] Secondly, embodiments of the present invention provide a meteorological risk early warning system, comprising:

[0011] processor;

[0012] Storage device, on which computer programs are stored,

[0013] When the computer program is executed by the processor, the processor implements any of the meteorological risk early warning methods described above for geological disaster analysis.

[0014] This invention provides a readable storage medium storing a program or instructions, which, when executed by a processor, implement the steps of the meteorological risk early warning method applied to geological disaster analysis.

[0015] This invention provides efficient and accurate early warning of geological disaster risks in target geological areas. Firstly, numerical meteorological change analysis is performed on meteorological element monitoring data to comprehensively and deeply uncover the patterns of change in these elements. Changes in meteorological elements are often a significant trigger for geological disasters; accurately grasping meteorological changes provides a crucial meteorological information foundation for subsequent geological disaster early warnings, enabling warnings to fully consider the dynamic impact of meteorological factors.

[0016] Among these methods, the fusion of meteorological change analysis results with geological radar detection information, combining GIS spatial analysis strategies for early warning processing, breaks through the limitations of independent analysis of meteorological and geological information in traditional methods. It fully utilizes the functions of GIS spatial analysis, organically combining the spatial distribution of meteorological elements with the vulnerability of geological structures. Through spatial interpolation, feature extraction, and overlay analysis, potential risk areas for geological disasters can be identified more accurately, improving the accuracy and targeting of early warnings.

[0017] Furthermore, a meteorological and geological risk early warning knowledge base was established based on meteorological element monitoring data, ground-penetrating radar detection information, and geological disaster risk early warning results. This knowledge base not only contains rich historical data and analysis results, but also reveals the causal relationship between meteorology, geology, and risk through association rule mining and knowledge graph construction. This enables rapid and accurate geological disaster risk assessment when faced with new real-time meteorological monitoring data, providing comprehensive and systematic knowledge support for geological disaster early warning and prevention.

[0018] Finally, the meteorological and geological risk early warning knowledge base is used to conduct geological disaster risk assessment on real-time meteorological monitoring data and output real-time geological disaster risk early warning information. This realizes the real-time and dynamic nature of geological disaster early warning, and can provide timely and effective decision-making basis by adjusting the warning level and scope in a timely manner according to the latest meteorological conditions, thereby minimizing the losses caused by geological disasters.

[0019] In summary, the embodiments of the present invention achieve deep integration of meteorological and geological information, effective accumulation and utilization of knowledge, and real-time and accurate geological disaster early warning. Attached Figure Description

[0020] Figure 1 This is a flowchart of a meteorological risk early warning method for geological disaster analysis provided in an embodiment of the present invention.

[0021] Figure 2 This is a schematic diagram of the basic structure of a meteorological risk early warning system provided in an embodiment of the present invention. Detailed Implementation

[0022] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0023] See Figure 1 As shown, this figure is a flowchart of a meteorological risk early warning method for geological disaster analysis provided by an embodiment of the present invention. This method can be executed by a meteorological risk early warning system. Figure 1 As shown, the method may include steps 110-140.

[0024] Step 110: Perform numerical meteorological change analysis on the meteorological element monitoring data of the target geological area to generate meteorological change analysis results.

[0025] For example, the target geological area contains all strata except for the Carboniferous, Cretaceous, and Tertiary periods. The Upper Proterozoic Sinian and Paleozoic strata are widely distributed, mainly consisting of littoral and shallow marine sediments interbedded with basic volcanic rocks, covering two-thirds of the total area of ​​the target geological area, with a total thickness of 6000m. The Mesozoic strata only expose the Triassic strata, which are clastic and carbonate rocks. The Triassic strata are widely distributed, covering one-third of the target geological area, with a thickness of approximately 4000m. The Cenozoic strata mainly consist of Quaternary alluvial, diluvial, debris flow, colluvial, and residual colluvial deposits, distributed along river valleys and in plains and sloping areas, with localized glacial deposits, reaching a maximum thickness of about 100m. It is evident that the target geological area has a relatively complex geological structure and relatively harsh meteorological conditions.

[0026] This invention aims to analyze meteorological element monitoring data of a target geological area to uncover meteorological change patterns and generate meteorological change analysis results. The meteorological element monitoring data includes real-time or historical observations of various meteorological elements within the target geological area, such as temperature, humidity, air pressure, wind speed, and wind direction. This data is collected from meteorological monitoring stations distributed throughout the target geological area, with each station recording meteorological element values ​​at corresponding time intervals.

[0027] When performing numerical weather change analysis on these meteorological element monitoring data, the data is first preprocessed. Preprocessing includes data cleaning to remove outliers and missing values ​​to ensure the accuracy and completeness of the data. For example, some monitoring stations may record meteorological element values ​​that deviate significantly from the normal range due to equipment failure or other reasons; these outliers will be identified and removed. Simultaneously, for missing data, appropriate methods will be used to fill in the gaps, such as interpolating values ​​using data from adjacent monitoring stations.

[0028] Next, various numerical analysis methods will be used to analyze the preprocessed data. Statistical characteristics of meteorological elements, such as mean, standard deviation, maximum, and minimum values, can be calculated to understand the overall distribution of these elements. The trends of meteorological elements over time will also be analyzed, such as determining whether temperature is trending upwards or downwards, and the rate of change of this trend. Furthermore, the correlations between different meteorological elements will be studied, such as whether there is a relationship between temperature and humidity. Through these analyses, a comprehensive understanding of the changes in meteorological elements within the target geological area can be obtained, generating meteorological change analysis results.

[0029] It should be understood that the target geological area can also be other areas. The embodiments of the present invention are only examples and are not intended to limit the target geological area.

[0030] Step 120: Combine the meteorological change analysis results with the geological radar detection information of the target geological area using GIS spatial analysis strategies to perform fusion early warning processing and obtain the geological disaster risk early warning results for the target geological area.

[0031] This invention integrates meteorological change analysis results with ground-penetrating radar (GPR) information, and utilizes GIS spatial analysis strategies to achieve early warning of geological disaster risks in target geological areas. Meteorological change analysis results reflect the changes in meteorological elements within the target geological area, while GPR information provides relevant information about the underground geological structure of the target area. Combining these two analyses allows for a more comprehensive and accurate assessment of geological disaster risks in the target geological area.

[0032] Step 121: Use GIS spatial analysis tools to perform spatial interpolation on the meteorological change analysis results to generate a spatial distribution map of meteorological elements in the target geological area.

[0033] In this step, GIS spatial analysis tools are needed to perform spatial interpolation on the meteorological change analysis results. Meteorological change analysis results are typically based on discrete meteorological monitoring station data, which are spatially discontinuous. To obtain the continuous spatial distribution of meteorological elements within the target geological area, spatial interpolation is required.

[0034] Spatial interpolation utilizes GIS spatial analysis tools. Based on the location of meteorological monitoring stations and their corresponding meteorological element values, appropriate interpolation algorithms are used to estimate the meteorological element values ​​at other locations within the target geological area. Exemplary interpolation algorithms include inverse distance weighted interpolation and Kriging interpolation. Inverse distance weighted interpolation determines the weight based on the distance between the point to be estimated and known monitoring stations; the closer the distance, the greater the weight. Kriging interpolation, on the other hand, considers the spatial correlation and variability of the data, enabling a more accurate estimation of values ​​at unknown points.

[0035] Spatial interpolation transforms discrete meteorological monitoring data into a continuous spatial distribution map of meteorological elements. This map displays the spatial distribution of meteorological elements within the target geological area, such as the specific values ​​of temperature and humidity at different locations. This spatial distribution map provides an intuitive spatial representation of meteorological information for subsequent integration with ground-penetrating radar (GPR) data.

[0036] Step 122: Extract geological structural features from the ground-penetrating radar information to obtain information on the distribution of geological structural vulnerability in the target geological area.

[0037] Ground-penetrating radar (GPR) information is obtained by detecting the underground geological structure of a target geological area using GPR equipment. This information includes data such as the reflected signals of underground geological layers and the propagation time of electromagnetic waves, reflecting the characteristics of the geological structure.

[0038] When extracting geological structural features from ground-penetrating radar (GPR) data, the raw data is first preprocessed. Preprocessing includes noise removal and signal strength enhancement to improve data quality. Then, using predefined algorithms and techniques, key geological structural features are extracted from the preprocessed data. These features include stratum thickness, lithology, and the morphology and orientation of geological structures.

[0039] By analyzing these geological structural characteristics, the vulnerability of geological structures within a target geological region can be assessed. Different geological structures exhibit varying stability and disaster resistance when facing external factors such as weather changes. For example, areas with thinner strata and looser lithology are more prone to geological disasters, exhibiting relatively higher geological structural vulnerability; while areas with thicker strata and harder lithology have relatively stable geological structures and lower vulnerability.

[0040] Based on the assessment results of geological structural characteristics and geological structural vulnerability, information on the distribution of geological structural vulnerability in the target geological area is generated. This information is presented in the form of a map or data table, showing the degree of geological structural vulnerability at different locations within the target geological area.

[0041] Step 123: Import the spatial distribution map of meteorological elements and the information on the distribution of geological structural vulnerability into the spatial overlay analysis network to perform multi-layer spatial correlation calculations and generate an initial risk warning layer.

[0042] After obtaining the spatial distribution maps of meteorological elements and the distribution information of geological structural vulnerability, these need to be imported into a spatial overlay analysis network for multi-layer spatial correlation calculations. The spatial overlay analysis network is an analysis model based on GIS technology that can overlay and analyze multiple spatial layers.

[0043] The spatial distribution map of meteorological elements and the distribution information of geological structural vulnerability represent meteorological and geological information within the target geological region, respectively, and exist in different layer formats. In the spatial overlay analysis network, these two layers are overlaid to match them spatially. Then, through spatial correlation operations, the relationship between meteorological elements and geological structural vulnerability is analyzed.

[0044] Spatial correlation calculations take into account the impact of meteorological factors on geological structures. In areas with high rainfall, if the geological structure is more vulnerable, the likelihood of geological disasters increases. Through a comprehensive analysis of meteorological factors and geological structural vulnerability, a risk value is assigned to each spatial location, reflecting the potential probability of geological disasters occurring at that location.

[0045] Based on the calculated risk values, an initial risk warning layer is generated. This layer uses different colors or symbols to represent different risk levels, visually displaying the geological hazard risk situation at various locations within the target geological area. The initial risk warning layer provides preliminary visualization results for subsequent risk assessment and early warning.

[0046] Step 124: Perform spatial clustering analysis on the initial risk warning layer to identify areas of abnormal risk clustering, calculate the risk impact range parameters of each cluster, and perform normalization processing on the risk impact range parameters to generate a dimensionless risk impact range index.

[0047] Performing spatial clustering analysis on the initial risk warning layer aims to further explore the characteristics of risk distribution and identify areas where risks are abnormally clustered. Spatial clustering analysis is a method of grouping spatially adjacent regions with similar risk characteristics.

[0048] In the initial risk warning layer, risk values ​​may vary at different locations. Some areas have relatively high risk values ​​and are clustered together, forming risk anomalous clusters. Spatial clustering analysis algorithms are used to group these areas with similar risk characteristics into the same cluster. Exemplary spatial clustering analysis algorithms include DBSCAN and K-Means.

[0049] After identifying areas of abnormal risk clustering, it is necessary to calculate the risk impact range parameters for each cluster. These parameters include the area, boundaries, and sum of risk values ​​of the clustered areas, reflecting the scale and severity of the abnormal risk clusters.

[0050] To facilitate comparison and comprehensive analysis between different clusters, the risk impact range parameter needs to be normalized. Normalization converts risk impact range parameters with different dimensions and value ranges into dimensionless values. This eliminates dimensional differences between parameters, allowing for comparison of the risk impact ranges of different clusters on the same scale. Based on the normalized results, a dimensionless risk impact range index is generated, which comprehensively reflects the degree of risk impact in each anomalous risk cluster.

[0051] Step 125: Based on the risk impact range index and the vulnerability level index in the geological structure vulnerability distribution information, generate a geological disaster risk early warning result that includes the risk level classification result through weighted fusion processing.

[0052] After obtaining the risk impact range index and vulnerability level indicators from the geological structural vulnerability distribution information, a weighted fusion process is needed to generate geological disaster risk early warning results. Weighted fusion processing is a method that comprehensively considers multiple factors, assigning different weights to different factors and fusing them to obtain more accurate results.

[0053] The risk impact range index reflects the scale and severity of areas with abnormal risk clustering, while the vulnerability level index in the geological structural vulnerability distribution information reflects the geological structural stability at different locations within the target geological region. In the weighted fusion processing, these two factors are assigned corresponding weights based on their respective impacts on geological hazard risk.

[0054] For areas with a large risk impact range index and a high level of geological structural vulnerability, the likelihood and severity of geological disasters are relatively high, resulting in a higher comprehensive risk value when weighted and integrated; while for areas with a small risk impact range index and a low level of geological structural vulnerability, the comprehensive risk value is relatively low.

[0055] Based on the weighted and fused comprehensive risk value, the target geological area is classified into risk levels. The comprehensive risk value is divided into different intervals, each interval corresponding to a risk level, such as low risk, medium risk, and high risk.

[0056] Finally, a geological disaster risk warning result containing the risk level classification is generated. This result is presented in the form of a map or report, accurately and comprehensively showing the geological disaster risk level of each location in the target geological area.

[0057] Step 130: Establish a meteorological and geological risk early warning knowledge base based on meteorological element monitoring data, geological radar detection information, and geological disaster risk early warning results.

[0058] Meteorological element monitoring data, ground-penetrating radar detection information, and geological disaster risk early warning results reflect relevant information about the target geological area from meteorological, geological, and risk assessment perspectives, respectively. Integrating this information to establish a meteorological and geological risk early warning knowledge base can provide comprehensive and systematic knowledge support for subsequent geological disaster risk assessment and early warning.

[0059] Step 131: Perform data standardization processing on the meteorological element monitoring data to generate a standardized meteorological dataset.

[0060] Meteorological element monitoring data come from different monitoring stations, and there may be differences in data format, units, and value ranges. To facilitate subsequent analysis and processing, it is necessary to perform data standardization processing on the meteorological element monitoring data.

[0061] Data standardization begins with unifying the data format. Different monitoring stations may use different data recording formats; through data conversion and organization, these formats are transformed into a unified format to ensure data consistency. Next, the dimensions of the data are processed. Meteorological elements have different dimensions; for example, temperature is measured in degrees Celsius, and wind speed in meters per second. To avoid the impact of dimensional differences on the analysis results, all meteorological element data are converted to the same dimension or subjected to dimensionless processing.

[0062] In addition, the range of data values ​​will be adjusted. The ranges of values ​​for different meteorological elements can vary significantly. Standardization methods unify the data range to a defined interval, such as between 0 and 1, allowing data from different meteorological elements to be compared and analyzed on the same scale. After data standardization, a standardized meteorological dataset is generated, containing the processed meteorological element monitoring data.

[0063] Step 132: Perform geological feature parameter extraction processing on the ground-penetrating radar detection information to obtain a set of geological structure feature parameters.

[0064] Ground-penetrating radar (GPR) detection information contains rich information about the underground geological structure of the target geological area, but the original detection data is quite complex, and it is necessary to extract key geological feature parameters from it.

[0065] When performing geological feature parameter extraction processing on ground-penetrating radar (GPR) data, the raw data is first parsed and processed. GPR data is typically stored in binary or text file format and needs to be converted into an analyzable format. Then, using specialized geological analysis algorithms and techniques, characteristic parameters of the geological structure are extracted from the data. These parameters include stratum thickness, lithology, morphology and strike of geological structures, and groundwater level. Stratum thickness reflects the vertical distribution of different geological layers, lithology determines the physical and mechanical properties of the geological body, the morphology and strike of geological structures affect the stability of the geological body, and changes in groundwater level also have a significant impact on the mechanical properties of the geological body. Through in-depth analysis and processing of GPR data, a set of geological structure characteristic parameters is obtained, which contains the key characteristic parameters of the geological structure within the target geological area.

[0066] Step 133: Import the standardized meteorological dataset, the set of geological structural feature parameters, and the geological disaster risk early warning results into the knowledge modeling algorithm, perform knowledge unit division processing, and generate multiple meteorological and geological knowledge units.

[0067] After obtaining standardized meteorological datasets, sets of geological structural characteristic parameters, and geological disaster risk early warning results, they need to be imported into knowledge modeling algorithms for knowledge unit partitioning. Knowledge modeling algorithms are methods that can analyze and process data, transforming it into knowledge units.

[0068] First, standardized meteorological datasets, sets of geological structural characteristic parameters, and geological hazard risk early warning results are integrated. These data are correlated and matched according to relevant rules to ensure spatial and temporal consistency. Then, using knowledge modeling algorithms, the integrated data is divided into multiple knowledge units based on its characteristics and inherent relationships. Each meteorological-geological knowledge unit contains a set of related meteorological, geological, and risk information. A knowledge unit may contain meteorological element values, geological structural characteristic parameters, and corresponding geological hazard risk levels for a target geological area within a specific time period. This knowledge unit division allows for the structuring and modularization of complex meteorological, geological, and risk information. After knowledge unit division, multiple meteorological-geological knowledge units are generated, providing the basic knowledge elements for constructing a meteorological-geological risk early warning knowledge base.

[0069] Step 134: Perform association rule mining on each meteorological and geological knowledge unit to identify the causal relationships between different knowledge units.

[0070] Step 1341: Extract the feature description vectors of each meteorological and geological knowledge unit and establish the feature matrix of the knowledge unit.

[0071] Meteorological and geological knowledge units contain information from various aspects such as meteorology, geology, and risk. In order to analyze and compare them, it is necessary to extract the feature description vectors of each meteorological and geological knowledge unit.

[0072] Feature description vectors are a quantitative representation of meteorological and geological knowledge units, containing numerical values ​​of key features within each unit. Statistical characteristics of meteorological elements, geological structural parameters, and geological hazard risk levels are extracted from these units and combined into vectors. Each vector represents a feature of a meteorological and geological knowledge unit. The feature description vectors of all meteorological and geological knowledge units are combined to construct a knowledge unit feature matrix. Each row of the matrix corresponds to a meteorological and geological knowledge unit, and each column corresponds to a feature. The knowledge unit feature matrix provides a visual representation of the feature distribution across all meteorological and geological knowledge units.

[0073] Step 1342: Calculate the feature similarity value between any two meteorological and geological knowledge units in the knowledge unit feature matrix.

[0074] After establishing the feature matrix of knowledge units, it is necessary to calculate the feature similarity value between any two meteorological and geological knowledge units. The feature similarity value reflects the degree of similarity between two meteorological and geological knowledge units, and is calculated by comparing their feature description vectors. Exemplary methods for calculating feature similarity values ​​include Euclidean distance and cosine similarity. Euclidean distance calculates the spatial distance between two feature description vectors; the closer the distance, the higher the similarity. Cosine similarity measures the similarity by calculating the cosine of the angle between two feature description vectors; the closer the cosine value is to 1, the higher the similarity. By calculating the feature similarity value between any two meteorological and geological knowledge units in the feature matrix, a similarity matrix can be obtained. Each element in the similarity matrix represents the feature similarity value between the corresponding two meteorological and geological knowledge units.

[0075] Step 1343: Construct a knowledge unit association network based on feature similarity values, and mark knowledge unit pairs with feature similarity values ​​greater than a preset similarity threshold as potential association unit pairs.

[0076] Based on the calculated feature similarity values, a knowledge unit association network is constructed. This network is a graph structure where each node represents a meteorological and geological knowledge unit, and edges between nodes indicate relationships between two units. A preset similarity threshold is set during network construction. When the feature similarity value of two meteorological and geological knowledge units exceeds this threshold, they are considered to have a potential relationship, and these two knowledge unit pairs are marked as potentially related. By marking potentially related pairs, the network highlights potentially related knowledge unit pairs, providing a selection range for subsequent causal relationship analysis.

[0077] Step 1344: Perform temporal causal relationship analysis on potential related unit pairs and extract the occurrence time series of potential related unit pairs in historical data.

[0078] For knowledge unit pairs marked as potentially related, temporal causal relationship analysis needs to be performed. Temporal causal relationship analysis is a method for studying the chronological order and causal relationship between two events. It involves extracting the time series of occurrences of potentially related unit pairs from historical data. The time series records the occurrence of each potentially related unit pair at different points in time. By analyzing the time series, the patterns and chronological order of occurrence of potentially related unit pairs can be understood. In the scenario of geological disaster analysis, it may be found that changes in certain meteorological conditions may lead to geological disasters after a period of time; temporal causal relationship analysis can identify this temporal causal relationship.

[0079] Step 1345: Calculate the lead-lag relationship parameters of potential related unit pairs based on time series data to determine the direction of causal association.

[0080] Based on the extracted time series, the lead-lag relationship parameters of potential related unit pairs are calculated. These parameters reflect the temporal order and time interval between two knowledge units. Through analysis and calculation of the time series, it is determined which knowledge unit appears first and which appears later in the potential related unit pair, as well as the time interval between them. If the appearance of one knowledge unit always precedes that of another, and a corresponding temporal pattern exists, then the former can be considered the cause and the latter the effect, thus determining the direction of the causal relationship. For example, in geological disaster analysis, if it is found that an increase in rainfall always precedes the occurrence of geological disasters, and a corresponding time interval exists, then the increase in rainfall can be considered a cause of the geological disasters, determining the direction of the causal relationship.

[0081] Step 1346: Calculate the causal association strength index based on the lead-lag relationship parameters and feature similarity values, and identify potential association unit pairs with causal association strength indices greater than the preset strength threshold as knowledge unit pairs with causal association relationships.

[0082] Based on the calculated lead-lag relationship parameters and feature similarity values, a causal association strength index is calculated. This index comprehensively considers the temporal sequence and feature similarity of two knowledge units, reflecting the tightness of their causal relationship.

[0083] By comprehensively analyzing the leading-lag relationship parameters and feature similarity values, a causal correlation strength index is obtained through a predetermined calculation method. Then, a preset strength threshold is set; when the causal correlation strength index of a potential correlated unit pair exceeds this threshold, they are considered to have a causal relationship, and these potential correlated unit pairs are identified as knowledge unit pairs with causal relationships. By identifying knowledge unit pairs with causal relationships, a clear causal correlation network can be established among meteorological and geological knowledge units.

[0084] Step 135: Construct a knowledge graph structure based on causal relationships, using meteorological and geological knowledge units as nodes and causal relationships as directed edges to generate the initial knowledge graph of the meteorological and geological risk early warning knowledge base.

[0085] After identifying the causal relationships between meteorological and geological knowledge units, a knowledge graph structure needs to be constructed based on these relationships. A knowledge graph is a graph-based knowledge representation method that can intuitively display the relationships between knowledge units. Meteorological and geological knowledge units are used as nodes in the knowledge graph, with each node representing a meteorological and geological knowledge unit and containing relevant information for that unit. Causal relationships are used as directed edges in the knowledge graph, with the direction of the directed edges indicating the direction of the causal relationship, from the causal knowledge unit to the result knowledge unit. By combining meteorological and geological knowledge units and causal relationships, an initial knowledge graph for the meteorological and geological risk early warning knowledge base is constructed. The initial knowledge graph graphically displays the causal relationship network between meteorological and geological knowledge units.

[0086] Step 136: Perform knowledge conflict detection processing on the initial knowledge graph to identify and correct contradictory relationships between knowledge units, and obtain the optimized meteorological and geological risk early warning knowledge base.

[0087] Step 1361: Traverse all directed edges in the initial knowledge graph and collect descriptions of causal relationships between knowledge unit pairs.

[0088] The directed edges in the initial knowledge graph represent the causal relationships between pairs of knowledge units. To identify and correct contradictory relationships, it is necessary to traverse all directed edges in the initial knowledge graph and collect descriptions of the causal relationships between each pair of knowledge units. Each directed edge is analyzed, recording the two knowledge units it connects to and the causal relationship information between them, such as the direction and strength of the causal relationship. By traversing all directed edges, a comprehensive understanding of the causal relationships between pairs of knowledge units in the initial knowledge graph can be obtained.

[0089] Step 1362: Perform a logical consistency check on the descriptions of multiple causal relationships between the same knowledge unit pairs to detect whether there are contradictory association directions or association strength indicators.

[0090] After collecting descriptions of causal relationships between pairs of knowledge units, it may be found that multiple causal relationship descriptions exist within the same knowledge unit pair. These descriptions may come from different analytical methods or data sources, requiring logical consistency checks. The checks should examine whether there are contradictory directions or strength indices in the causal relationship descriptions within the same knowledge unit pair. In some cases, one description may state that knowledge unit A is the cause of knowledge unit B, while another description states that knowledge unit B is the cause of knowledge unit A, i.e., a contradiction in the direction of the relationship; or two descriptions may have significantly different assessments of the strength of the causal relationship, i.e., a contradiction in the strength indices.

[0091] By verifying logical consistency, these contradictory relationships can be discovered, providing a basis for subsequent corrections.

[0092] Step 1363: When a contradictory relationship is detected, extract the actual frequency of occurrence and probability of association of the knowledge unit in historical data.

[0093] When contradictory relationships are detected, further analysis of the actual situation of the knowledge unit pair in historical data is needed. The actual frequency of occurrence and the probability of association between the knowledge unit pair are extracted from the historical data. The actual frequency of occurrence records the number of times the knowledge unit pair appears simultaneously in the historical data, while the probability of association is the frequency with which the knowledge unit pair exhibits a correlation. By analyzing the actual frequency of occurrence and the probability of association, the true causal relationship between the knowledge unit pairs can be understood.

[0094] Step 1364: Recalculate the causal association strength index based on the actual frequency of occurrence and the probability of association occurrence, and retain the association descriptions with higher causal association strength indices.

[0095] Based on the extracted actual frequency of occurrence and probability of association, the causal association strength index is recalculated. The actual frequency of occurrence and probability of association reflect the stability and reliability of the causal relationships between knowledge unit pairs. Incorporating these into the calculation of the causal association strength index yields more accurate results. By comparing the causal association strength indices of different relationship descriptions between the same knowledge unit pairs, descriptions with higher causal association strength indices are retained. This removes potentially inaccurate or unreliable relationship descriptions, improving the accuracy of the knowledge graph.

[0096] Step 1365: Perform secondary association rule mining on the corrected knowledge units to eliminate contradictory associations.

[0097] A secondary association rule mining process is performed on the corrected knowledge unit pairs. This process, based on the corrected knowledge unit pairs, further delves into and analyzes the relationships between them to ensure that contradictory relationships are completely eliminated. Using association rule mining algorithms, the corrected knowledge unit pairs are re-analyzed to identify more accurate relationships between them. Through this secondary association rule mining process, the relationships between knowledge unit pairs can be further optimized, improving the quality of the knowledge graph.

[0098] Step 1366: Update the corrected knowledge unit relationships to the initial knowledge graph to generate an optimized meteorological and geological risk early warning knowledge base.

[0099] The knowledge unit relationships, after being revised and processed through secondary association rule mining, are updated to the initial knowledge graph. The updated knowledge graph eliminates contradictory relationships and more accurately reflects the causal relationships between knowledge units. Finally, an optimized meteorological and geological risk early warning knowledge base is generated. The optimized meteorological and geological risk early warning knowledge base contains accurate and reliable meteorological, geological, and risk knowledge.

[0100] Step 140: Use the meteorological and geological risk early warning knowledge base to perform geological disaster risk assessment on the real-time meteorological monitoring data of the target geological area, and output the real-time geological disaster risk early warning information of the target geological area.

[0101] The meteorological and geological risk early warning knowledge base integrates knowledge from various sources, including meteorological element monitoring data, geological radar detection information, and geological disaster risk early warning results. It encompasses the causal relationships between meteorology, geology, and risk. By analyzing and evaluating real-time meteorological monitoring data of target geological areas using this knowledge base, potential geological disaster risks can be identified promptly, and real-time geological disaster risk early warning information can be output.

[0102] Step 141: Perform structured parsing processing on the real-time meteorological monitoring data to extract a set of real-time meteorological characteristic parameters that include meteorological element types, monitoring periods, and element change trends.

[0103] Real-time meteorological monitoring data is meteorological information of a target geological area collected in real time by meteorological monitoring equipment. This data usually exists in raw form and needs to be structured and analyzed to extract useful information.

[0104] When performing structured analysis on real-time meteorological monitoring data, the first step is to identify and classify the data based on its format. Different types of meteorological element data are separated, such as temperature, humidity, air pressure, and wind speed. Then, key information for each meteorological element is extracted, including the element type, monitoring period, and its changing trend.

[0105] The meteorological element type clarifies which meteorological element is being monitored, the monitoring period records the time range for data collection, and the element change trend reflects how the meteorological element changes within the monitoring period—whether it is increasing, decreasing, or remaining stable. Combining this information generates a real-time set of meteorological characteristic parameters that includes the meteorological element type, monitoring period, and element change trend.

[0106] Step 142: Input the set of real-time meteorological feature parameters into the dynamic association layer of the meteorological and geological risk early warning knowledge base, perform knowledge unit matching processing based on the hierarchical semantic relationship of the terminology dictionary, and establish dynamic knowledge association results between real-time meteorological features and historical meteorological and geological knowledge units in the knowledge base.

[0107] Step 1421: Extract meteorological characteristic description terms from historical meteorological and geological knowledge units in the meteorological and geological risk early warning knowledge base, and construct a hierarchical term index tree. The hierarchical term index tree includes meteorological element type branches, element change trend branches, and related geological characteristic branches.

[0108] The meteorological and geological risk early warning knowledge base contains a large number of historical meteorological and geological knowledge units, each with corresponding meteorological characteristic descriptive terms. Meteorological characteristic descriptive terms are extracted from these knowledge units to construct a hierarchical terminology index tree.

[0109] A hierarchical terminology index tree is a tree-like structure that includes branches for meteorological element types, element change trends, and related geological features. The meteorological element type branch contains terms for various meteorological elements, such as temperature, humidity, and wind speed; the element change trend branch records terms for the changing trends of meteorological elements, such as rising, falling, and stable; and the related geological feature branch contains terms for geological features related to meteorological elements, such as stratigraphy and lithology, and geological structures. By constructing a hierarchical terminology index tree, the meteorological feature description terms of historical meteorological and geological knowledge units can be structured and organized.

[0110] Step 1422: Perform terminology normalization on the meteorological element type terms in the real-time meteorological feature parameter set, correct synonymous terms with different names by referring to the standard terminology system of the hierarchical terminology index tree, and generate a standardized meteorological terminology set.

[0111] The meteorological element type terms in the real-time meteorological characteristic parameter set may have synonyms but different names. In order to ensure the consistency and accuracy of terminology, it is necessary to perform terminology normalization processing on these terms.

[0112] Referring to the standard terminology system of the hierarchical terminology index tree, the terminology for meteorological element types in the real-time meteorological characteristic parameter set is revised. Synonymous terms with different names are unified into standard terms, such as unifying "air temperature" as "temperature".

[0113] A standardized meteorological terminology set is generated through terminology normalization. The terms in the standardized meteorological terminology set are consistent with the standard terminology system of the hierarchical terminology index tree, providing an accurate terminology basis for subsequent knowledge unit matching.

[0114] Step 1423: Using the standardized meteorological terminology set as the search keyword, traverse the corresponding branch nodes of the hierarchical terminology index tree and extract historical meteorological and geological knowledge units containing terms of the same meteorological element type as candidate knowledge unit sets.

[0115] Using a standardized set of meteorological terms as search keywords, a search is conducted within a hierarchical terminology index tree. The meteorological element type branches of the hierarchical terminology index tree are traversed to find historical meteorological and geological knowledge units containing terms of the same meteorological element type.

[0116] Historical meteorological and geological knowledge units containing terms of the same meteorological element type were extracted and used as a candidate knowledge unit set. This candidate knowledge unit set, selected from historical meteorological and geological knowledge units, is a set of knowledge units related to the meteorological element types in the real-time meteorological characteristic parameter set, providing a narrower scope for further matching.

[0117] Step 1424: Perform context element matching processing on the candidate knowledge unit set, and compare the semantic similarity between the description of the change trend of elements in the real-time meteorological feature parameter set and the description of the change trend of elements in each unit of the candidate knowledge unit set. The semantic similarity is determined by the path overlap of terms in the hierarchical term index tree and the number of matching context-related terms.

[0118] The context element matching process is performed on the candidate knowledge unit set, which mainly compares the semantic similarity between the description of the element change trend in the real-time meteorological feature parameter set and the description of the element change trend of each unit in the candidate knowledge unit set.

[0119] Semantic similarity is determined by the path overlap of terms in the hierarchical term index tree and the number of matches between context-related terms. In the hierarchical term index tree, each term has its defined path; the higher the path overlap, the more semantically similar the two terms are. Simultaneously, the number of matches between context-related terms also reflects the relevance of the two descriptions.

[0120] If the temperature trend is described as "rising" in the real-time meteorological feature parameter set, and the temperature trend is also described as "rising" in a certain knowledge unit in the candidate knowledge unit set, and their paths overlap significantly in the hierarchical term index tree, and the number of matched context-related terms is also large, then the semantic similarity between the two descriptions is high.

[0121] Step 1425: Mark the candidate knowledge units whose semantic similarity meets the preset matching threshold as associated knowledge units. Combine all associated knowledge units and their matching point descriptions with real-time meteorological features to generate dynamic knowledge association results. The dynamic knowledge association results include associated knowledge unit identifiers, matching term sets, and semantic similarity descriptions.

[0122] A preset matching threshold is set. When the semantic similarity between a candidate knowledge unit in the candidate knowledge unit set and the description of the change trend of elements in the real-time meteorological feature parameter set meets the threshold, the candidate knowledge unit is marked as an associated knowledge unit.

[0123] All knowledge units marked as related knowledge units, along with their matching point descriptions with real-time meteorological characteristics, are organized and recorded. The matching point descriptions record the aspects in which the related knowledge units match the real-time meteorological characteristics, such as meteorological element types and element change trends.

[0124] Finally, dynamic knowledge association results are generated. These results include identifiers of associated knowledge units, sets of matching terms, and semantic similarity descriptions, establishing dynamic knowledge associations between real-time meteorological features and historical meteorological and geological knowledge units in the knowledge base.

[0125] Step 143: Call the hybrid reasoning layer of the meteorological and geological risk early warning knowledge base, combine the dynamic knowledge association results to construct a multi-dimensional risk transmission path, identify the initial transmission node through rule reasoning, and verify the effectiveness of node association through case reasoning, and generate a multi-dimensional risk transmission path containing key triggering factors and transmission intensity descriptions.

[0126] Step 1431: Extract the historical risk transmission rules corresponding to the associated knowledge units from the dynamic knowledge association results. The historical risk transmission rules include the premise meteorological element terms, the intermediate geological feature terms, and the conclusion disaster type terms.

[0127] The dynamic knowledge association results include historical meteorological and geological knowledge units associated with real-time meteorological characteristics. Corresponding historical risk transmission rules are extracted from these associated knowledge units. These rules describe the causal relationships between meteorological elements, geological features, and geological hazards. They include: prerequisite meteorological element terms (i.e., the meteorological conditions leading to geological hazards); intermediate geological feature terms (reflecting changes in geological structure under the influence of meteorological elements); and concluding hazard type terms (indicating the possible types of geological hazards). By extracting these historical risk transmission rules, we can understand how geological hazards occur and develop under similar meteorological conditions.

[0128] Step 1432: Starting with the standardized meteorological terms in the real-time meteorological feature parameter set, match the premise meteorological element terms of the historical risk transmission rules, and perform rule reasoning to obtain the initial transmission node. The initial transmission node contains intermediate geological feature terms and corresponding rule confidence descriptions.

[0129] Starting with standardized meteorological terms from the real-time meteorological feature parameter set, a matching process is performed within historical risk transmission rules. Rules in the historical risk transmission rules that have the same or similar prerequisite meteorological element terms as the standardized meteorological terms are searched. Once a matching rule is found, rule reasoning is performed. Based on the logical relationships of the rules, intermediate geological feature terms are derived from the prerequisite meteorological element terms, resulting in the initial transmission node. The initial transmission node contains the intermediate geological feature terms and the corresponding rule confidence description. The rule confidence description reflects the reliability and accuracy of the rule. Obtaining the initial transmission node through rule reasoning determines the starting point and intermediate links for constructing a multi-dimensional risk transmission path.

[0130] Step 1433: Based on the initial transmission nodes, traverse the knowledge graph of the meteorological and geological risk early warning knowledge base, extract downstream geological feature terms and disaster type terms that are directly related to intermediate geological feature terms, and construct the initial risk transmission path graph.

[0131] Based on the obtained initial transmission nodes, the knowledge graph of the meteorological and geological risk early warning knowledge base is traversed. The knowledge graph records the causal relationships between meteorological and geological knowledge units.

[0132] Starting from the intermediate geological feature terms of the initial transmission node, we search for downstream geological feature terms and hazard type terms that are directly related to it. These downstream geological feature terms and hazard type terms reflect the further development and possible consequences of geological hazards in the geological structure.

[0133] By connecting initial transmission nodes, downstream geological feature terms, and disaster type terms, an initial risk transmission path map is constructed. This initial risk transmission path map illustrates the preliminary risk transmission path from meteorological elements to geological features and then to geological disasters.

[0134] Step 1434: Call the case reasoning module of the hybrid reasoning layer, select the historical meteorological and geological knowledge unit with the highest correlation from the dynamic knowledge association results, and extract its complete risk transmission path as the verification case path.

[0135] The system utilizes the case-based reasoning module within the hybrid reasoning layer of the meteorological and geological risk early warning knowledge base. This module is a tool for reasoning based on historical cases.

[0136] Historical meteorological and geological knowledge units with the highest correlation were selected from the dynamic knowledge association results. These units, which are most similar to real-time meteorological characteristics, contain information that has high reference value for current risk assessments.

[0137] Complete risk transmission paths were extracted from selected historical meteorological and geological knowledge units to serve as verification case paths. These verification case paths documented the actual occurrence and development of geological disasters under similar meteorological and geological conditions.

[0138] Step 1435: Compare the overlap between the initial risk transmission path map and the execution nodes of the verification case path, retain the path branches whose overlap meets the preset overlap, and delete the path branches whose overlap is lower than the preset overlap to obtain the initial optimized transmission path.

[0139] The initial risk transmission path map is compared with the validation case path to determine the degree of node overlap. Node overlap comparison involves comparing the number and location of identical nodes in the two paths.

[0140] Set a preset overlap degree. When the overlap degree between a node of a path branch in the initial risk transmission path map and a node of the verification case path meets the preset overlap degree, the path branch is retained; when the node overlap degree is lower than the preset overlap degree, the path branch is deleted.

[0141] By comparing and filtering node overlap, path branches that do not conform to the actual case are removed, resulting in an initial optimized transmission path. This initial optimized transmission path is more consistent with reality, improving the accuracy of the risk transmission path.

[0142] Step 1436: Perform iterative verification processing on the initial optimized transmission path. Starting from each node in the path, rematch historical risk transmission rules and verification case paths, correct the association descriptions between nodes, supplement missing intermediate transmission nodes, and generate a multi-dimensional risk transmission path containing node terminology sequences, node association rule descriptions, and case verification results. The transmission strength description is determined by combining the confidence of node association rules and the overlap of case verification.

[0143] The initial optimized transmission path undergoes iterative verification. Starting from each node in the initial optimized transmission path, it is matched again against historical risk transmission rules to find relevant rules. Simultaneously, it is compared again with the verification case path.

[0144] During the iterative verification process, the descriptions of relationships between nodes are revised. It may be discovered that the relationships between some nodes are inaccurate in real-world cases and require correction. Missing intermediate transmission nodes are also added. Some risk transmission processes may have intermediate steps not reflected in the initial optimized transmission path; iterative verification can identify and supplement these intermediate transmission nodes.

[0145] Finally, a multi-dimensional risk transmission path is generated, comprising a sequence of node terms, descriptions of node association rules, and case validation results. The transmission strength description is determined by combining the confidence level of the node association rules and the overlap of case validation. The confidence level of the node association rules reflects the reliability of the rules, while the overlap of case validation reflects the degree to which the path conforms to actual cases. Taking them together yields a more accurate description of the transmission strength.

[0146] Step 144: Based on the multi-dimensional risk transmission path and the risk level association rule base in the meteorological and geological risk early warning knowledge base, perform risk level logical judgment processing, match the risk rule triggering conditions corresponding to each node in the path, and determine the real-time geological disaster risk level through rule priority sorting and trigger frequency statistics.

[0147] Step 1441: Analyze the risk level association rule base in the meteorological and geological risk early warning knowledge base, and extract the rule set corresponding to each risk level. The rule set includes triggering conditions such as the combination of triggering elements, the length of the transmission path, and the description of the node association strength.

[0148] The risk level association rule base in the meteorological and geological risk early warning knowledge base contains rule sets corresponding to different risk levels. The risk level association rule base is parsed to extract the rule sets corresponding to each risk level. Each rule set contains a series of triggering conditions, such as combinations of triggering elements, transmission path length, and node association strength descriptions. Triggering element combinations specify which combinations of meteorological and geological elements will trigger that risk level; transmission path length reflects the complexity of risk transmission; and node association strength describes the closeness of the association between nodes in the risk transmission path. By parsing the risk level association rule base, the triggering conditions for different risk levels are clarified.

[0149] Step 1442: Traverse each node term in the multi-dimensional risk transmission path, match the trigger element combination conditions of the corresponding node in the risk level association rule base, and record the rule entries that meet the conditions and the corresponding risk levels.

[0150] In this embodiment of the invention, node terms are the core semantic units constituting the multi-dimensional risk transmission path in the meteorological and geological risk early warning knowledge base, covering three core terminology branches: meteorological elements, geological features, and disaster types. For example, the meteorological element category includes standardized meteorological terms such as temperature, humidity, and rainfall; the geological feature category includes geological parameter terms such as stratigraphic lithology and geological structural fragility; and the disaster type category involves geological disaster terms such as landslides and debris flows. It can be understood that node terms must meet the semantic specifications of a hierarchical terminology index tree, exist as inference nodes in the risk transmission path, and verify the transmission relationship between nodes through association rules and case inference. Their semantic matching degree and association strength directly affect the risk level determination result.

[0151] Iterate through each node term in the multi-dimensional risk transmission path. Search the risk level association rule base for the triggering element combinations corresponding to these node terms. When a node term in the multi-dimensional risk transmission path satisfies the triggering element combination conditions of a rule in the risk level association rule base, record the rule entry and its corresponding risk level. Through traversal and matching, find all rule entries in the multi-dimensional risk transmission path that satisfy the triggering conditions in the risk level association rule base and their corresponding risk levels.

[0152] Step 1443: Perform rule priority sorting on the recorded rule entries. Sort the rules from high to low according to the preset priority in the risk level association rule base. The priority sorting is determined based on the number of nodes in the transmission path covered by the rule and the trigger success rate in historical cases.

[0153] The recorded rule entries are sorted by rule priority. In the risk level association rule base, each rule has a preset priority. Priority is determined based on the number of nodes in the transmission path covered by the rule and the trigger success rate in historical cases. The more nodes in the transmission path covered by a rule, the more comprehensive its description of the risk transmission process; the higher the trigger success rate in historical cases, the stronger the rule's reliability. The rule entries are sorted from highest to lowest according to the preset priority.

[0154] Step 1444: Calculate the frequency of rule triggering for each risk level, assign weighted statistical coefficients to the rules with higher priority, and generate weighted trigger frequency statistics.

[0155] The system calculates the rule trigger frequency for each risk level. Rule trigger frequency refers to the number of times a rule corresponding to a specific risk level is triggered in a multi-dimensional risk transmission path. Rules with higher priority are assigned weighted statistical coefficients. These coefficients highlight the importance of higher-priority rules; the higher the priority, the larger the weighted coefficient. The rule trigger frequency is then multiplied by its corresponding weighted coefficient to obtain the weighted trigger frequency. The weighted trigger frequencies for each risk level are summed to generate the weighted trigger frequency statistics. This weighted trigger frequency statistical result comprehensively considers both rule priority and trigger frequency, more accurately reflecting the probability of each risk level.

[0156] Step 1445: Select the risk level with the highest frequency in the weighted trigger frequency statistics as the candidate risk level, and determine whether the rule entry corresponding to the candidate risk level covers the key trigger factor nodes in the multi-dimensional risk transmission path. If it covers, the candidate risk level is determined as the real-time geological disaster risk level. If it does not cover, the second highest frequency risk level is selected for re-verification until a real-time geological disaster risk level that meets the coverage condition is determined.

[0157] In the weighted trigger frequency statistics, the risk level with the highest frequency is selected as the candidate risk level. Then, it is determined whether the rule entries corresponding to the candidate risk level cover the key trigger factor nodes in the multi-dimensional risk transmission path.

[0158] Key triggering factor nodes are nodes that play a crucial role in the occurrence of geological disasters within the multi-dimensional risk transmission path. If the rule entries corresponding to a candidate risk level cover these key triggering factor nodes, it indicates that the risk level can accurately reflect the current geological disaster risk situation, and the candidate risk level is determined as the real-time geological disaster risk level.

[0159] If the rule entries corresponding to the candidate risk level do not cover the key triggering factor nodes, the second highest frequency risk level in the weighted triggering frequency statistics is selected as the new candidate risk level, and the verification is repeated until a real-time geological disaster risk level that meets the coverage conditions is found.

[0160] Step 145: Combine real-time geological disaster risk levels, geographical knowledge units covered by multi-dimensional risk transmission paths, and temporal correlation patterns in historical cases to generate real-time geological disaster risk warning information that includes risk level descriptions, sets of knowledge units of affected areas, and estimated occurrence periods.

[0161] Step 1451: Extract the descriptions of geographical knowledge units covered by the multi-dimensional risk transmission path from the meteorological and geological risk early warning knowledge base. The geographical knowledge unit descriptions include administrative division terms, topographic feature terms, and key geographical element terms.

[0162] The multi-dimensional risk transmission path covers several geographical locations within the target geological area, and the corresponding geographical information is stored in the meteorological and geological risk early warning knowledge base. Descriptions of geographical knowledge units covered by the multi-dimensional risk transmission path are extracted from this knowledge base. These descriptions include administrative division terms, clarifying the administrative region to which the area belongs; topographic feature terms, describing the topographical characteristics of the area, such as mountains and plains; and key geographical element terms, indicating important geographical elements within the area, such as rivers and mountain ranges. By extracting these geographical knowledge unit descriptions, the geographical scope and characteristics affected by the multi-dimensional risk transmission path can be further explored.

[0163] Step 1452: Map the description of the geographic knowledge unit to the preset geographic coding system to generate a set of influence area knowledge units containing unique identifiers of the geographic knowledge units.

[0164] The pre-defined geocoding system is a system for encoding geographic regions, with each region having a unique code. Geographic knowledge unit descriptions are mapped to this pre-defined geocoding system. Based on the administrative division terms, topographic feature terms, and key geographic element terms in the geographic knowledge unit descriptions, the corresponding codes are found in the geocoding system. These codes are combined to generate a set of impact area knowledge units containing unique identifiers for the geographic knowledge units. This set of impact area knowledge units clearly defines the geographic regions affected by multi-dimensional risk transmission paths.

[0165] Step 1453: Extract the temporal characteristic description of the historical meteorological and geological knowledge unit with the highest correlation in the dynamic knowledge association results.

[0166] The dynamic knowledge association results include historical meteorological and geological knowledge units associated with real-time meteorological characteristics. The historical meteorological and geological knowledge unit with the highest correlation is selected from these units. The temporal feature description of this historical meteorological and geological knowledge unit is then extracted. The temporal feature description records the time information of the meteorological and geological events corresponding to this knowledge unit, such as the specific time of occurrence and duration of the event. By extracting the temporal feature description, we can understand the temporal patterns of geological disasters under similar meteorological and geological conditions.

[0167] Step 1454: Referring to the time feature description and the monitoring start time in the real-time meteorological feature parameter set, perform time correlation pattern matching processing. By semantically matching the time interval term of meteorological element change-disaster occurrence in historical cases with the real-time monitoring duration term, determine the time range description of the estimated occurrence period.

[0168] Referring to the extracted time feature descriptions and the monitoring start times in the real-time meteorological feature parameter set, time correlation pattern matching processing is performed. Historical cases record the time intervals between changes in meteorological elements and the occurrence of geological disasters. Semantic matching is performed between the terminology of meteorological element change-disaster occurrence time intervals in historical cases and the terminology of real-time monitoring duration. If the semantic match is successful, it indicates that the current meteorological conditions and historical cases are similar, and the occurrence time of geological disasters can be estimated based on the time intervals in historical cases. Based on the matching results, a time range description for the estimated occurrence period is determined. The time range description for the estimated occurrence period provides the possible time interval for geological disasters, providing time-dimensional information for early warning information.

[0169] Step 1455: Organize the textual description of the real-time geological disaster risk level, the geocoded list of the knowledge unit set of the affected area, and the time range description of the estimated occurrence period into structured data, supplement the knowledge unit references of the corresponding risk level in the meteorological and geological risk early warning knowledge base, and generate real-time geological disaster risk early warning information that includes a risk overview, details of the affected area, potential occurrence time range, and reference identifiers of the response measures.

[0170] The textual descriptions of real-time geological disaster risk levels, the geocoded list of knowledge units representing affected areas, and the time range descriptions of estimated occurrence periods are compiled into structured data. Corresponding response knowledge units for each risk level are retrieved from the meteorological and geological risk early warning knowledge base. These response knowledge units contain methods and suggestions for responding to geological disasters of different risk levels. The structured data and response knowledge unit references are combined to generate real-time geological disaster risk early warning information that includes a risk overview, details of the affected area, the potential time range of occurrence, and response reference identifiers.

[0171] As a non-limiting embodiment, the method further includes:

[0172] Step 210: Collect historical geological disaster case data, meteorological observation data, geological survey reports and academic research literature of the target geological area to construct a multi-source knowledge material library.

[0173] To further improve the meteorological and geological risk early warning knowledge base, it is necessary to collect multi-source knowledge materials for the target geological area, including historical geological disaster case data, meteorological observation data, geological survey reports, and academic research literature.

[0174] Historical geological disaster case data records detailed information about past geological disasters within the target geological area, such as the time, location, type, scale, and damage. Meteorological observation data includes observations of meteorological elements within the target geological area at different time periods, such as temperature, humidity, air pressure, and wind speed. Geological survey reports are detailed reports obtained after surveying the underground geological structure of the target geological area, including information on stratigraphy, lithology, geological structure, and hydrogeology. Academic research literature comprises the research results of experts and scholars in related fields on geological disaster issues in the target geological area, containing new theories, methods, and research findings. These multi-source knowledge materials are collected and organized to construct a multi-source knowledge resource library. This multi-source knowledge resource library provides a rich source of information for subsequent knowledge extraction and integration.

[0175] Step 220: Perform natural language processing on the text data in the multi-source knowledge material library to extract entity terms and relational terms related to geological disasters.

[0176] Text data in multi-source knowledge repositories contains a large amount of textual information, requiring natural language processing (NLP) techniques to extract useful information. When performing NLP on text data, the first step is text preprocessing. This includes removing noise such as punctuation and stop words, and segmenting the text into individual words. Then, named entity recognition (NER) technology is used to extract geological hazard-related entity terms from the preprocessed text. Entity terms include geological hazard types (e.g., landslides, debris flows, earthquakes); meteorological elements (e.g., temperature, humidity, rainfall); and geological structural features (e.g., stratigraphy, geological structure). Simultaneously, relation extraction techniques are used to identify relationships between entities. Relationship terms describe causal, inclusion, and associative relationships between entities. For example, a text might mention "increased rainfall led to landslides," and relation extraction techniques can identify a causal relationship between "rainfall" and "landslide." Through NLP, geological hazard-related entity and relation terms are extracted.

[0177] Step 230: Perform standardization processing on entity terms, establish a unified terminology dictionary, and merge the same entities with different expressions into standard entities.

[0178] Extracted entity terms may have different expressions. To facilitate subsequent knowledge management and application, standardization is necessary. Standardization begins with classifying and organizing the terms. Entity terms of the same type are grouped together, such as grouping all terms representing geological hazard types into one category. Then, a unified terminology dictionary is established. This dictionary contains standard expressions for all entity terms. Through comparison and analysis, identical entities with different expressions are merged into standard entities. For example, in different texts, "landslide" and "landslide" might be used to represent the same geological hazard; these are merged into the standard entity "landslide". Standardization and the establishment of the terminology dictionary ensure the consistency and accuracy of entity terms, providing a standardized terminology foundation for subsequent knowledge fusion and map construction.

[0179] Step 240: Perform semantic parsing on relational terms to identify causal, inclusion, and association relationships between entities and construct triplet knowledge units.

[0180] Semantic parsing of relational terms is performed to gain a deeper understanding of the relationships between entities. Semantic parsing analyzes the meaning of relational terms and identifies causal, inclusion, and associative relationships between entities. A causal relationship indicates that a change in one entity leads to a change in another entity. For example, increased rainfall leads to landslides; there is a causal relationship between "increased rainfall" and "landslide occurrence." An inclusion relationship indicates that one entity contains another entity. Strata may contain different lithologies; there is an inclusion relationship between "strata" and "lithology." An associative relationship indicates that there is some connection between two entities, but it is not necessarily a causal or inclusion relationship. Temperature and humidity may simultaneously affect the occurrence of geological hazards; there is an associative relationship between "temperature" and "humidity." Based on the identified entities and relationships, triplet knowledge units are constructed. A triplet knowledge unit consists of entity 1, a relationship, and entity 2, such as (increased rainfall leads to, landslide occurrence). By constructing triplet knowledge units, the relationships between entities are represented in a structured form, providing basic knowledge units for subsequent knowledge graph construction.

[0181] Step 250: Merge the ternary knowledge units with the current knowledge units in the meteorological and geological risk early warning knowledge base, establish connections through entity linking technology, and supplement the nodes and edges of the knowledge graph.

[0182] The meteorological and geological risk early warning knowledge base already contains corresponding knowledge units and a knowledge graph. The newly constructed triplet knowledge units are then integrated with the existing knowledge units in the meteorological and geological risk early warning knowledge base. Entity linking technology is used to match and associate entities in the triplet knowledge units with entities in the meteorological and geological risk early warning knowledge base. If an entity in the triplet knowledge unit is the same as or similar to an entity in the meteorological and geological risk early warning knowledge base, they are linked. During the integration process, the relationships within the triplet knowledge units are added as new edges to the knowledge graph, and entities are added as new nodes to supplement the knowledge graph. This approach supplements the nodes and edges of the knowledge graph, making it more complete and richer.

[0183] Step 260: Perform knowledge completion processing on the fused knowledge graph. Based on the existing triplet knowledge units and preset reasoning rules, predict the missing relationships and generate the completed knowledge graph.

[0184] The merged knowledge graph may contain some missing relationships. To improve it, knowledge completion processing is required. Based on existing triplet knowledge units and pre-defined inference rules, knowledge completion is performed. These pre-defined inference rules are summarized from knowledge and experience in the field of geological hazards and are used to infer potential relationships between entities. Given that "increased rainfall" leads to "landslides," and "landslides" lead to "road damage," the inference rules can infer a possible indirect causal relationship between "increased rainfall" and "road damage." Through knowledge completion processing, missing relationships are predicted and added to the merged knowledge graph, generating a completed knowledge graph. The completed knowledge graph more comprehensively reflects the relationships between entities.

[0185] Step 270: Perform a quality assessment on the completed knowledge graph and correct any erroneous knowledge units and relationships.

[0186] The completed knowledge graph may contain some erroneous knowledge units and relationships, requiring a quality assessment. This assessment examines the completed knowledge graph from multiple perspectives: checking the accuracy of knowledge units to ensure correct representation of entities and relationships; and checking the rationality of relationships to determine if the connections between entities align with the realities of the geological disaster field. Incorrect causal relationships may exist in the knowledge graph, such as incorrectly associating "temperature rise" as the cause of "earthquake occurrence." This error can be detected through the quality assessment. When erroneous knowledge units and relationships are found, they will be corrected. Erroneous relationships will be deleted, and inaccurate knowledge units will be updated to improve the quality of the knowledge graph.

[0187] Step 280: Update the knowledge graph that has passed the quality assessment to the meteorological and geological risk early warning knowledge base.

[0188] The quality of the knowledge graph, after quality assessment and revision, has been guaranteed. The knowledge graph that passed the quality assessment has been updated into the meteorological and geological risk early warning knowledge base. The updated knowledge base contains richer and more accurate knowledge, providing stronger knowledge support for the analysis and early warning of geological disasters. Continuous updating and improvement of the knowledge base can enhance the accuracy and reliability of geological disaster risk assessment and early warning.

[0189] In a scalable embodiment, the method further includes:

[0190] Step 310: Based on the geological structural vulnerability distribution information and historical geological disaster case data in the meteorological geological risk early warning knowledge base, construct a three-dimensional geological model of the target geological area. The model includes key geological parameters such as stratigraphy, geological structure, topography, and hydrogeology.

[0191] The geological vulnerability distribution information and historical geological disaster case data in the meteorological and geological risk early warning knowledge base provide detailed information on the geological conditions of the target geological area. Based on this information, a three-dimensional geological model of the target geological area is constructed.

[0192] When constructing a three-dimensional geological model, the first step is to collect key geological parameters such as stratigraphic lithology, geological structure, topography, and hydrogeology from the information on the distribution of geological structural vulnerabilities. Stratigraphic lithology determines the physical and mechanical properties of geological bodies, geological structure reflects the morphology and orientation of underground geological structures, topography affects the flow and collection of surface water, and hydrogeology involves the distribution and movement of groundwater.

[0193] These key geological parameters are integrated and visualized using 3D modeling software. Different colors and textures are used to represent different lithologies; underground geological structure models are constructed based on the orientation of geological structures; surface topographic models are generated by combining topographic and geomorphological data; and hydrogeological information is incorporated into the model to simulate groundwater flow.

[0194] By constructing a three-dimensional geological model, the geological structure and characteristics of the target geological area can be displayed intuitively.

[0195] Step 320: Input real-time meteorological monitoring data into the three-dimensional geological model to simulate the physical and mechanical response of the geological body under different meteorological conditions. The physical and mechanical response includes changes in pore water pressure, decrease in rock and soil strength, and changes in slope stability.

[0196] Real-time meteorological monitoring data reflects the current meteorological conditions of the target geological area. This data is then input into a three-dimensional geological model to simulate the physical and mechanical responses of geological bodies under different meteorological conditions.

[0197] Changes in meteorological conditions can have multifaceted effects on geological formations. Increased rainfall can lead to a rise in groundwater levels, thereby increasing pore water pressure; temperature changes can affect the thermal expansion and contraction of soil and rock, resulting in changes in the strength of the soil and rock; changes in wind speed and direction can exert additional forces on slopes, affecting their stability.

[0198] In a three-dimensional geological model, the physical and mechanical responses of geological bodies are simulated based on parameters such as rainfall, temperature, and wind speed from real-time meteorological monitoring data. Numerical simulation methods are used to calculate changes in pore water pressure, the attenuation of soil and rock strength, and slope stability.

[0199] Increased pore water pressure reduces the effective stress of soil and rock, thereby reducing their strength. This decrease in strength increases the likelihood of slope instability, and changes in slope stability are directly related to the occurrence of geological disasters.

[0200] By simulating the physical and mechanical responses of geological bodies under different meteorological conditions, we can understand the dynamic changes of geological bodies under the influence of meteorological changes.

[0201] Step 330: Set various meteorological scenario parameters, including rainfall, rainfall intensity, temperature change and wind force level, and perform simulation calculations in the three-dimensional geological model to generate prediction results of the probability of geological disaster occurrence and the scope of impact under different scenarios.

[0202] To comprehensively assess the likelihood and impact of geological disasters under different meteorological conditions, it is necessary to set various meteorological scenario parameters. These parameters include rainfall, rainfall intensity, temperature variation, and wind speed.

[0203] In the three-dimensional geological model, different combinations of meteorological scenario parameters are set. These include scenarios with high rainfall, high rainfall intensity, rising temperature, and strong winds, as well as scenarios with low rainfall, low rainfall intensity, falling temperature, and light winds.

[0204] For each combination of meteorological scenario parameters, simulations are performed in a three-dimensional geological model. The simulations consider the physical and mechanical responses of the geological body to meteorological conditions, as well as the relationship between these responses and the occurrence of geological hazards. Through numerical simulation and analysis, the probability of geological hazards occurring under different meteorological scenarios is calculated.

[0205] Under conditions of heavy and intense rainfall, slope stability decreases due to increased pore water pressure and reduced soil and rock strength, thus increasing the probability of geological disasters. The predicted impact range of geological disasters is also crucial. This range is related to factors such as the physical and mechanical properties of the geological body, topography, and meteorological conditions. In areas with steep terrain and loose soil and rock, the impact range of geological disasters may be even wider.

[0206] Finally, the prediction results of the probability of occurrence and the scope of impact of geological disasters under different scenarios are generated. These results are presented in the form of data tables or maps, which intuitively show the probability of occurrence of geological disasters and the areas that may be affected under different meteorological conditions.

[0207] Step 340: Perform sensitivity analysis on the simulation results to identify the meteorological parameters that have the greatest impact on the probability of geological disasters and determine the key influencing factors.

[0208] Sensitivity analysis was performed on the simulation results under different scenarios. Sensitivity analysis is a method to study the degree to which changes in input parameters affect the output results. In geological disaster simulation, the input parameters are meteorological scenario parameters, and the output results are the probability of geological disaster occurrence.

[0209] By changing the value of a certain meteorological parameter while keeping other meteorological parameters constant, the changes in the probability of geological disasters are observed. When rainfall changes, the magnitude of the change in the probability of geological disasters is observed; when temperature changes, the changes in the probability of geological disasters are observed again.

[0210] Based on the degree of influence of changes in meteorological parameters on the probability of geological disasters, the meteorological parameters with the greatest impact on the probability of geological disasters are identified. If a small change in rainfall leads to a significant change in the probability of geological disasters, then rainfall is a key influencing factor.

[0211] Sensitivity analysis was used to identify key influencing factors. Identifying these key influencing factors helps in geological disaster early warning and prevention, allowing for a focus on meteorological parameters that significantly impact the probability of geological disasters, thereby improving the accuracy of early warnings and the targeted nature of preventative measures.

[0212] Step 350: Based on the sensitivity analysis results and the changing trends of real-time meteorological monitoring data, dynamically adjust the input parameters of the three-dimensional geological model to perform dynamic simulation of the geological disaster development process.

[0213] Based on the sensitivity analysis results, the meteorological parameters that have the greatest impact on the probability of geological disasters were identified. Meanwhile, real-time meteorological monitoring data reflects the changing trends of current meteorological conditions in the target geological area.

[0214] The input parameters of the three-dimensional geological model are dynamically adjusted based on the results of sensitivity analysis and the changing trends of real-time meteorological monitoring data. If the sensitivity analysis results indicate that rainfall is a key influencing factor, and real-time meteorological monitoring data shows that rainfall is increasing, then the value of the rainfall parameter in the three-dimensional geological model is increased accordingly.

[0215] After dynamically adjusting the input parameters, a dynamic simulation of the geological hazard development process is performed in a 3D geological model. The dynamic simulation measures the entire process of a geological hazard from its occurrence to its development over time. Under conditions of continuously increasing rainfall, it simulates the gradual instability of slopes, and the gradual formation and development of landslides or debris flows.

[0216] By dynamically predicting the development process of geological disasters, we can understand their development trends in real time, take corresponding preventive measures in advance, and reduce the losses caused by geological disasters.

[0217] Step 360: Compare and verify the dynamic simulation results with real-time geological disaster risk early warning information, and correct the parameter settings of the three-dimensional geological model.

[0218] The dynamic simulation results of the geological disaster development process are compared and verified with real-time geological disaster risk early warning information. The real-time geological disaster risk early warning information is obtained by analyzing and evaluating real-time meteorological monitoring data through a meteorological and geological risk early warning knowledge base, and includes information such as real-time geological disaster risk level, affected area, and estimated occurrence time.

[0219] Compare the risk level, affected area, and estimated occurrence time in the dynamic simulation results with the real-time geological disaster risk warning information. Check whether the probability of occurrence, scope of impact, and occurrence time of geological disasters in the dynamic simulation results are consistent with the real-time geological disaster risk warning information.

[0220] If discrepancies are found between the two, the parameter settings of the 3D geological model need to be corrected. If the dynamic simulation results show a wider impact range of geological hazards than the real-time geological hazard risk warning information, this may be due to unreasonable parameter settings in the 3D geological model, and these parameters need to be adjusted.

[0221] By comparing and verifying and correcting parameters, the accuracy and reliability of three-dimensional geological models can be improved, making the simulation and early warning of geological disasters more in line with reality.

[0222] Step 370: Combine the corrected 3D geological model with parameter settings to generate a geological disaster simulation report that includes simulation results of different scenarios, sensitivity analysis reports, and dynamic simulation animations.

[0223] After refining the parameter settings of the 3D geological model, a geological hazard simulation report is generated based on the revised 3D geological model. The geological hazard simulation report includes simulation results for different scenarios, a sensitivity analysis report, and dynamic simulation animations.

[0224] Simulation results under different scenarios demonstrate the predicted probability and impact range of geological disasters under various meteorological conditions. These results are presented in the form of charts and maps, intuitively reflecting the influence of meteorological conditions on geological disasters. The sensitivity analysis report details the meteorological parameters that have the greatest impact on the probability of geological disasters, i.e., the key influencing factors. The report analyzes the degree of influence of each meteorological parameter on the probability of geological disasters and provides corresponding analytical conclusions.

[0225] Dynamic simulation animations visually demonstrate the entire process of geological disasters from occurrence to development. The animations accurately and comprehensively show the physical and mechanical responses of geological bodies under the influence of meteorological conditions, as well as the formation and development process of geological disasters. Through dynamic simulation animations, one can more intuitively understand the development trend and potential impacts of geological disasters.

[0226] By integrating simulation results from different scenarios, sensitivity analysis reports, and dynamic simulation animations, a geological disaster simulation report is generated. This report provides comprehensive and detailed information for early warning, prevention, and decision-making regarding geological disasters, and effective measures can be taken based on the report to reduce losses caused by geological disasters.

[0227] See Figure 2 As shown in the figure, this is a schematic diagram of the basic structure of a meteorological risk warning system 200 provided in an embodiment of the present invention. The meteorological risk warning system 200 includes:

[0228] Processor 201;

[0229] Storage device 202, on which computer program 2020 is stored;

[0230] When the computer program 2020 is executed by the processor 201, the processor 201 implements any of the meteorological risk early warning methods applied to geological disaster analysis.

[0231] Based on the above, a readable storage medium is provided, on which a program or instructions are stored, and when the program or instructions are executed by a processor, the steps of the above method are implemented.

[0232] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems or apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple, and relevant parts can be referred to the method section.

Claims

1. A meteorological risk early warning method applied to geological disaster analysis, characterized in that, The meteorological risk early warning method applied to geological disaster analysis includes: Numerical meteorological change analysis is performed on meteorological element monitoring data of the target geological area to generate meteorological change analysis results. By combining GIS spatial analysis strategies with meteorological change analysis results and ground-penetrating radar (GPR) detection information of the target geological area, a geological disaster risk warning result for the target geological area is obtained. Specifically, GIS spatial analysis tools are used to perform spatial interpolation on the meteorological change analysis results to generate a spatial distribution map of meteorological elements in the target geological area; geological structural features are extracted from the GPR detection information to obtain the geological structural vulnerability distribution information of the target geological area; the spatial distribution map of meteorological elements and the geological structural vulnerability distribution information are imported into a spatial overlay analysis network to perform multi-layer spatial correlation operations, generating an initial risk warning layer; spatial clustering analysis is performed on the initial risk warning layer to identify areas of abnormal risk clustering, calculate the risk impact range parameters of each cluster, and normalize the risk impact range parameters to generate a dimensionless risk impact range index; based on the risk impact range index and the vulnerability level index in the geological structural vulnerability distribution information, a geological disaster risk warning result containing risk level classification results is generated through weighted fusion processing. Establish a meteorological and geological risk early warning knowledge base based on meteorological element monitoring data, geological radar detection information, and geological disaster risk early warning results; The meteorological and geological risk early warning knowledge base is used to conduct geological disaster risk assessment on the real-time meteorological monitoring data of the target geological area and output real-time geological disaster risk early warning information for the target geological area.

2. The meteorological risk early warning method for geological disaster analysis as described in claim 1, characterized in that, The meteorological and geological risk early warning knowledge base established based on meteorological element monitoring data, ground-penetrating radar detection information, and geological disaster risk early warning results includes: Perform data standardization processing on meteorological element monitoring data to generate standardized meteorological datasets; Geological feature parameter extraction processing is performed on the information detected by ground-penetrating radar to obtain a set of geological structure feature parameters; The standardized meteorological dataset, the set of geological structural feature parameters, and the geological disaster risk early warning results are imported into the knowledge modeling algorithm, and knowledge unit division processing is performed to generate multiple meteorological and geological knowledge units. Perform association rule mining on each meteorological and geological knowledge unit to identify causal relationships between different knowledge units; A knowledge graph structure is constructed based on causal relationships, with meteorological and geological knowledge units as nodes and causal relationships as directed edges, generating the initial knowledge graph of the meteorological and geological risk early warning knowledge base. The initial knowledge graph is subjected to knowledge conflict detection processing to identify and correct contradictory relationships between knowledge units, resulting in an optimized meteorological and geological risk early warning knowledge base.

3. The meteorological risk early warning method for geological disaster analysis as described in claim 2, characterized in that, The process of performing association rule mining on each meteorological and geological knowledge unit to identify causal relationships between different knowledge units includes: Extract the feature description vectors of each meteorological and geological knowledge unit, and establish the feature matrix of the knowledge unit; Calculate the feature similarity value between any two meteorological and geological knowledge units in the knowledge unit feature matrix; A knowledge unit association network is constructed based on feature similarity values, and knowledge unit pairs with feature similarity values ​​greater than a preset similarity threshold are marked as potential association unit pairs; Analyze the causal relationship of potential related unit pairs in the execution time sequence and extract the time series of occurrence of potential related unit pairs in historical data; Based on time series analysis, the lead-lag relationship parameters of potential related unit pairs are calculated to determine the direction of causal association. The causal association strength index is calculated based on the lead-lag relationship parameters and feature similarity values. Potentially related unit pairs with causal association strength indices greater than a preset strength threshold are identified as knowledge unit pairs with causal relationships.

4. The meteorological risk early warning method for geological disaster analysis as described in claim 2, characterized in that, The process of performing knowledge conflict detection on the initial knowledge graph to identify and correct contradictory relationships between knowledge units yields an optimized meteorological and geological risk early warning knowledge base, including: Traverse all directed edges in the initial knowledge graph and collect descriptions of causal relationships between knowledge unit pairs; For multiple causal relationships between the same knowledge unit, perform logical consistency checks to detect whether there are contradictory association directions or association strength indicators. When a contradictory relationship is detected, extract the actual frequency of occurrence and the probability of the relationship in historical data for that knowledge unit; The causal association strength index is recalculated based on the actual frequency of occurrence and the probability of association occurrence, and the association description with a higher causal association strength index is retained. The corrected knowledge units are subjected to secondary association rule mining to eliminate contradictory associations; The revised knowledge unit relationships are updated to the initial knowledge graph to generate an optimized meteorological and geological risk early warning knowledge base.

5. The meteorological risk early warning method for geological disaster analysis as described in claim 1, characterized in that, The process involves using a meteorological and geological risk early warning knowledge base to perform geological hazard risk assessment on real-time meteorological monitoring data of the target geological area, and outputting real-time geological hazard risk early warning information for the target geological area, including: The real-time meteorological monitoring data is subjected to structured parsing processing to extract a set of real-time meteorological characteristic parameters that include meteorological element types, monitoring periods, and element change trends; The set of real-time meteorological feature parameters is input into the dynamic association layer of the meteorological and geological risk early warning knowledge base. Based on the hierarchical semantic relationship of the terminology dictionary, knowledge unit matching processing is performed to establish the dynamic knowledge association result between real-time meteorological features and historical meteorological and geological knowledge units in the knowledge base. The hybrid reasoning layer of the meteorological and geological risk early warning knowledge base is invoked, and multi-dimensional risk transmission path construction is performed by combining dynamic knowledge association results. Initial transmission nodes are identified through rule reasoning, and the effectiveness of node association is verified through case reasoning, generating multi-dimensional risk transmission paths that include key triggering factors and descriptions of transmission intensity. Based on the risk level association rule base in the multi-dimensional risk transmission path and meteorological and geological risk early warning knowledge base, the risk level logic judgment process is executed, the risk rule triggering conditions corresponding to each node in the path are matched, and the real-time geological disaster risk level is determined by the rule priority sorting and trigger frequency statistics. By combining real-time geological disaster risk levels, geographical knowledge units covered by multi-dimensional risk transmission paths, and temporal correlation patterns in historical cases, real-time geological disaster risk warning information is generated, which includes risk level descriptions, sets of knowledge units of affected areas, and estimated occurrence periods.

6. The meteorological risk early warning method for geological disaster analysis as described in claim 5, characterized in that, The process involves inputting a set of real-time meteorological feature parameters into the dynamic association layer of the meteorological and geological risk early warning knowledge base, performing knowledge unit matching processing based on the hierarchical semantic relationships of the terminology dictionary, and establishing a dynamic knowledge association result between real-time meteorological features and historical meteorological and geological knowledge units in the knowledge base, including: Meteorological characteristic description terms are extracted from historical meteorological and geological knowledge units in the meteorological and geological risk early warning knowledge base, and a hierarchical term index tree is constructed. The hierarchical term index tree includes meteorological element type branches, element change trend branches, and related geological characteristic branches. The meteorological element type terms in the real-time meteorological characteristic parameter set are normalized. Synonymous terms with different names are corrected with reference to the standard terminology system of the hierarchical terminology index tree, and a standardized meteorological terminology set is generated. Using the standardized meteorological terminology set as the search keyword, the corresponding branch nodes of the hierarchical terminology index tree are traversed to extract historical meteorological and geological knowledge units containing terms of the same meteorological element type as candidate knowledge unit sets. Context element matching processing is performed on the candidate knowledge unit set to compare the semantic similarity between the description of the change trend of elements in the real-time meteorological feature parameter set and the description of the change trend of elements in each unit of the candidate knowledge unit set. The semantic similarity is determined by the path overlap of terms in the hierarchical term index tree and the number of matched context-related terms. Candidate knowledge units whose semantic similarity meets a preset matching threshold are marked as associated knowledge units. Dynamic knowledge association results are generated by combining all associated knowledge units and their matching point descriptions with real-time meteorological features. The dynamic knowledge association results include associated knowledge unit identifiers, matching term sets, and semantic similarity descriptions. The hybrid reasoning layer, which invokes the meteorological and geological risk early warning knowledge base, combines dynamic knowledge association results to construct a multi-dimensional risk transmission path. It identifies initial transmission nodes through rule-based reasoning and verifies the effectiveness of node associations through case-based reasoning, generating a multi-dimensional risk transmission path that includes key triggering factors and descriptions of transmission intensity. This includes: The historical risk transmission rules corresponding to the associated knowledge units are extracted from the dynamic knowledge association results. The historical risk transmission rules include the premise meteorological element terms, intermediate geological feature terms, and conclusion disaster type terms. Starting with standardized meteorological terms in the real-time meteorological feature parameter set, the initial transmission node is obtained by matching the premise meteorological element terms of historical risk transmission rules and performing rule reasoning. The initial transmission node includes intermediate geological feature terms and corresponding rule confidence descriptions. Based on the knowledge graph of the meteorological and geological risk early warning knowledge base traversed by the initial transmission nodes, downstream geological feature terms and disaster type terms directly associated with intermediate geological feature terms were extracted to construct the initial risk transmission path graph. The case reasoning module of the hybrid reasoning layer is invoked to select the historical meteorological and geological knowledge unit with the highest correlation from the dynamic knowledge association results, and extract its complete risk transmission path as the verification case path. The overlap between the initial risk transmission path map and the execution nodes of the verification case path is compared. Path branches with an overlap that meets the preset overlap are retained, and path branches with an overlap that is lower than the preset overlap are deleted to obtain the initial optimized transmission path. The initial optimized transmission path is subjected to iterative verification. Starting from each node in the path, the historical risk transmission rules and verification case paths are rematched, the association descriptions between nodes are corrected, missing intermediate transmission nodes are supplemented, and a multi-dimensional risk transmission path containing node terminology sequences, node association rule descriptions and case verification results is generated. The transmission strength description is determined by combining the confidence of node association rules and the overlap of case verification.

7. The meteorological risk early warning method for geological disaster analysis as described in claim 5, characterized in that, The risk level association rule base based on the multi-dimensional risk transmission path and the meteorological and geological risk early warning knowledge base performs risk level logical judgment processing, matches the risk rule triggering conditions corresponding to each node in the path, and determines the real-time geological disaster risk level through rule priority sorting and trigger frequency statistics, including: The risk level association rule base in the meteorological and geological risk early warning knowledge base is analyzed, and the rule set corresponding to each risk level is extracted. The rule set includes triggering conditions such as the combination of triggering elements, the length of the transmission path and the description of the node association strength. Traverse each node term in the multi-dimensional risk transmission path, match the trigger element combination conditions of the corresponding node in the risk level association rule base, and record the rule entries that meet the conditions and their corresponding risk levels; The recorded rule entries are sorted by priority according to their preset priority in the risk level association rule base. The priority is determined based on the number of nodes in the transmission path covered by the rule and the trigger success rate in historical cases. The frequency of rule triggering for each risk level is statistically analyzed, and weighted statistical coefficients are assigned to the rules with higher priority to generate weighted trigger frequency statistics. The risk level with the highest frequency in the weighted trigger frequency statistics is taken as the candidate risk level. It is determined whether the rule entries corresponding to the candidate risk level cover the key trigger factor nodes in the multi-dimensional risk transmission path. If they cover, the candidate risk level is determined as the real-time geological disaster risk level. If they do not cover, the second highest frequency risk level is selected for re-verification until a real-time geological disaster risk level that meets the coverage condition is determined. The system combines real-time geological disaster risk levels, geographical knowledge units covering multi-dimensional risk transmission paths, and temporal correlation patterns in historical cases to generate real-time geological disaster risk early warning information, including risk level descriptions, sets of knowledge units of affected areas, and estimated occurrence periods. The description of geographical knowledge units covering the multi-dimensional risk transmission path is extracted from the meteorological and geological risk early warning knowledge base. The geographical knowledge unit description includes administrative division terms, topographic feature terms, and key geographical element terms. The descriptions of geographical knowledge units are mapped to a preset geographic coding system to generate a set of influence area knowledge units containing unique identifiers of geographical knowledge units; Extract the temporal characteristic description of the historical meteorological and geological knowledge unit with the highest correlation in the dynamic knowledge association results; Referring to the time feature description and the monitoring start time in the real-time meteorological feature parameter set, time correlation pattern matching processing is performed. By semantic matching of the time interval term of meteorological element change-disaster occurrence in historical cases with the real-time monitoring duration term, the time range description of the estimated occurrence period is determined. The textual description of real-time geological disaster risk levels, the geocoded list of knowledge units of affected areas, and the time range description of the estimated occurrence period are organized into structured data. The corresponding risk level response measures knowledge units in the meteorological and geological risk early warning knowledge base are supplemented to generate real-time geological disaster risk early warning information that includes a risk overview, details of the affected area, potential occurrence time range, and response measure reference identifiers.

8. A meteorological risk early warning system, characterized in that, include: processor; A storage device having a computer program stored thereon, which, when executed by the processor, causes the processor to implement the meteorological risk early warning method for geological disaster analysis as described in any one of claims 1-7.

9. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions, which, when executed by a processor, implement the meteorological risk early warning method for geological disaster analysis as described in any one of claims 1-7.

Images