Natural earthquake risk analysis method and system based on artificial neural network, and medium

Abstract

The invention discloses a natural earthquake risk analysis method based on an artificial neural network, and the method comprises the steps: S1, obtaining a target analysis region, and dividing the target analysis region into a plurality of sub-regions; s2, acquiring a historical seismic data set corresponding to the target analysis area and preprocessing the historical seismic data set; s3, performing sub-region marking on the preprocessed data set to obtain a plurality of seismic source sub-regions and non-seismic source sub-regions; s4, acquiring data subsets corresponding to the seismic source sub-region and the non-seismic source sub-region, and inputting the data subsets into an artificial neural network for feature extraction; s5, calculating the similarity between the non-seismic source sub-region and the plurality of seismic source sub-regions; s6, training the convolutional neural network by using the historical seismic data set to obtain a seismic prediction model; and S7, inputting the data subsets of the different regions into the earthquake prediction model for prediction to obtain an earthquake prediction result, and outputting an earthquake risk analysis report according to the earthquake prediction result. According to the method, earthquake risk analysis of multiple areas and potential earthquake areas is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of seismic analysis, and particularly to a method, system, device and medium for natural seismic hazard analysis based on an artificial neural network. Background Art

[0002] In the prior art, for the analysis of seismic hazards, a traditional probabilistic seismic hazard analysis method is generally adopted. The traditional probabilistic seismic hazard analysis method generally analyzes a single region, successively analyzes the influence of all seismic sources on this region, and then synthesizes the influence of each seismic source on this region to obtain the exceedance probability of the hazard of this region.

[0003] The above analysis of seismic hazards is carried out for a single region, and the result of each analysis is the result of a single region. It is impossible to obtain the seismic hazard analysis parameters of multiple regions simultaneously. If one wants to achieve the seismic hazard analysis of a multi-region system, a large amount of data calculation and processing processes need to be faced, and the time consumed is relatively long. Summary of the Invention

[0004] The purpose of the present invention is to provide a method, system, device and medium for natural seismic hazard analysis based on an artificial neural network. Through efficient data processing by the artificial neural network, it can simultaneously perform hazard analysis on multiple regions, and can adopt different monitoring strategies for different region types to achieve personalized analysis.

[0005] The present invention is realized by the following technical solutions:

[0006] In a first aspect, a method for natural seismic hazard analysis based on an artificial neural network provided by an embodiment of the present invention includes:

[0007] S1: Obtain a target analysis region and divide the target analysis region into multiple sub-regions;

[0008] S2: Obtain the historical earthquake data set corresponding to the target analysis region and perform preprocessing to obtain the preprocessed historical earthquake data set. The historical earthquake data set includes region data and earthquake data;

[0009] S3: Perform sub-region marking on the preprocessed historical earthquake data set to obtain multiple source sub-regions and multiple non-source sub-regions;

[0010] S4: Obtain the data subsets corresponding to the source sub-regions and the non-source sub-regions, input the data subsets into the artificial neural network for feature extraction, and obtain the feature vectors of the source sub-regions and the feature vectors of the non-source sub-regions;

[0011] S5: Calculate the similarity between the non-seismic source sub-regions and multiple seismic source sub-regions, mark the non-seismic source sub-regions with similarity greater than the preset threshold as potential seismic source sub-regions, and mark the non-seismic source sub-regions with similarity less than or equal to the preset threshold as non-seismic source sub-regions;

[0012] S6: Train the convolutional neural network using the historical earthquake dataset to obtain an earthquake prediction model;

[0013] S7: Adopt different monitoring strategies for the seismic source sub-regions, potential seismic source sub-regions, and non-seismic source sub-regions, collect the corresponding regional data subsets from the seismic source sub-regions, potential seismic source sub-regions, and non-seismic source sub-regions according to the monitoring strategies, input the regional data subsets into the earthquake prediction model for prediction to obtain earthquake prediction results, and output an earthquake risk analysis report based on the earthquake prediction results.

[0014] Further, the specific method for obtaining and preprocessing the historical earthquake dataset corresponding to the target analysis region includes:

[0015] Remove the abnormal data points in the historical earthquake dataset using the point-by-point elimination method to obtain the remaining data, and fill in the missing data in the remaining data.

[0016] Further, S21: Assume that X = {x0, x1, x2,..., x n} is the interpolation sample set, Y = {y0, y1, y2,..., y n} is the corresponding function values of X, n represents the number of nodes, and x represents the interpolation node;

[0017] S22: Input n + 1 interpolation sample points x0, x1, x2,..., x n and the corresponding function values y0, y1, y2,..., y n ;

[0018] S23: Calculate the nth-order Lagrange basis function:

[0019]

[0020] where i = 0, 1, 2,... n;

[0021] S24: Calculate the nth-order Lagrange interpolation function:

[0022]

[0023] where l i (x) represents the nth-order Lagrange basis function;

[0024] S25: Input the interpolation point x and substitute it into the interpolation function in step S24 to obtain the calculation result.

[0025] Furthermore, the specific method for marking sub-regions based on the pre-processed historical earthquake dataset includes:

[0026] Determine whether the location information of the earthquake source in the historical earthquake dataset is within a certain sub-region. If so, mark this sub-region as the earthquake source sub-region; if not, mark the sub-region as a non-earthquake source sub-region.

[0027] Furthermore, obtaining the data subsets corresponding to the earthquake source sub-regions and non-earthquake source sub-regions, and inputting the data subsets into an artificial neural network for feature extraction specifically includes:

[0028] Obtain the earthquake source dataset corresponding to the earthquake source sub-region and the non-earthquake source dataset corresponding to the non-earthquake source sub-region from the historical earthquake dataset. The earthquake source dataset includes the earthquake data subset and the earthquake source region data subset, and the non-earthquake source dataset includes the non-earthquake source region data subset;

[0029] Input the earthquake data subset and the earthquake source region data subset into the artificial neural network for feature extraction respectively to obtain the earthquake feature vectors and earthquake source region feature vectors corresponding to each earthquake source sub-region;

[0030] Input the non-earthquake source region data subset into the artificial neural network for feature extraction to obtain the non-earthquake source region feature vectors corresponding to each non-earthquake source sub-region.

[0031] Furthermore, training a convolutional neural network using the historical earthquake dataset to obtain an earthquake prediction model specifically includes:

[0032] Construct a training sample set by combining the earthquake source dataset and the non-earthquake source dataset according to a preset ratio, use the region data subset in the remaining earthquake source dataset as the test set, and use the earthquake data subset as the validation set;

[0033] Use the convolutional neural network as the model to be trained, input the training sample set into the convolutional neural network for training. The validation set is used to observe whether the loss function of the convolutional neural network converges during the training process to determine whether to terminate the training, and the test set is used to test the prediction accuracy of the convolutional neural network; Train the convolutional neural network using the stochastic gradient descent method until the loss function of the neural network converges to the minimum value, and use the trained convolutional neural network as the earthquake prediction model.

[0034] Furthermore, the monitoring strategy is to collect the regional data within the sub-region at different acquisition frequencies. The acquisition frequency of the earthquake source sub-region is greater than that of the potential earthquake source sub-region, and the acquisition frequency of the potential earthquake source sub-region is greater than that of the non-earthquake source sub-region.

[0035] Second aspect, a natural earthquake hazard analysis system based on an artificial neural network provided by another embodiment of the present invention is used to implement the natural earthquake hazard analysis method based on an artificial neural network described in the above embodiment, and includes:

[0036] A regional division module, configured to obtain a target analysis region and divide the target analysis region into multiple sub-regions;

[0037] A data processing module, configured to obtain a historical earthquake data set corresponding to the analysis region and perform preprocessing on it to obtain a preprocessed historical earthquake data set, where the historical earthquake data set includes regional data and earthquake data;

[0038] A regional marking module, configured to perform sub-region marking on the preprocessed historical earthquake data set to obtain multiple source sub-regions and multiple non-source sub-regions, and is further configured to calculate the similarity between the non-source sub-regions and the multiple source sub-regions, mark the non-source sub-regions with a similarity greater than a preset threshold as potential source sub-regions, and mark the non-source sub-regions with a similarity less than or equal to the preset threshold as non-source sub-regions without earthquakes;

[0039] A feature extraction module, configured to obtain data subsets corresponding to the source sub-regions and non-source sub-regions, input the data subsets into an artificial neural network for feature extraction, and obtain feature vectors of the source sub-regions and feature vectors of the non-source sub-regions;

[0040] A model training module, configured to train a convolutional neural network based on the historical earthquake data set to obtain an earthquake prediction model;

[0041] A regional monitoring module, configured to adopt different monitoring strategies for the source sub-regions, potential source sub-regions, and non-source sub-regions without earthquakes, collect corresponding regional data subsets from the source sub-regions, potential source sub-regions, and non-source sub-regions without earthquakes according to the monitoring strategies, input the regional data subsets into the earthquake prediction model for prediction to obtain an earthquake prediction result, and output an earthquake hazard analysis report according to the earthquake prediction result.

[0042] Third aspect, a natural earthquake hazard analysis device provided by another embodiment of the present invention, the device includes: a memory and a processor, the memory and the processor are coupled; the memory stores program instructions, and when the program instructions are executed by the processor, the device executes the method described in the above embodiment.

[0043] Fourth aspect, a computer-readable storage medium provided by another embodiment of the present invention includes a computer program, and when the computer program runs on an electronic device, the electronic device executes the method described in the above embodiment.

[0044] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0045] A natural earthquake hazard analysis method based on an artificial neural network provided by an embodiment of the present invention realizes earthquake hazard analysis for multiple regions. By accurately extracting the characteristics of the earthquake source region through the artificial neural network, potential earthquake regions in areas where earthquakes have not occurred are identified, effectively reducing the earthquake hazard in potential earthquake regions. In addition, earthquake prediction is carried out through a convolutional neural network to form an earthquake hazard analysis report, providing a scientific basis for managers to make earthquake prevention decisions and further reducing the harm caused by earthquakes. In addition, by adopting different monitoring strategies, the embodiment of the present invention realizes personalized monitoring of different types of regions, improving the fineness of earthquake hazard analysis.

[0046] The embodiment of the present invention conducts model training based on deep learning, reducing the calculation cost, being able to efficiently utilize historical data resources, and improving the efficiency of natural earthquake hazard analysis.

[0047] A natural earthquake hazard analysis system, device, and medium based on an artificial neural network provided by an embodiment of the present invention have the same inventive concept as the natural earthquake hazard analysis method based on an artificial neural network and have the same beneficial effects, which will not be elaborated herein. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] To more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the drawings required for the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts. In the drawings:

[0049] Figure 1 is a flowchart of a natural earthquake hazard analysis method based on an artificial neural network provided by the first embodiment of the present invention;

[0050] Figure 2 is a structural block diagram of a natural earthquake hazard analysis system based on an artificial neural network provided by the second embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0051] To make the objectives, technical solutions, and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the embodiments and the drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and do not limit the present invention.

[0052] Embodiment 1

[0053] AsFigure 1 As shown in Figure 1 , a natural earthquake hazard analysis method based on an artificial neural network provided by the first embodiment of the present invention includes the following steps:

[0054] S1: Obtain the target analysis area and divide the target analysis area into multiple sub-areas.

[0055] The area division method can be carried out according to administrative regions. For example, if the preset analysis area is one or more provinces, each sub-area can correspond to a city or district; if the preset analysis area is one or more countries, each sub-area can correspond to a province or city; or, an artificial division method can also be adopted to divide the preset analysis area into multiple sub-areas according to geographical location or geographical area. The area of each sub-area can be the same or different; or, any other known area division method in the art can also be adopted, and the present application does not make specific limitations on this.

[0056] S2: Obtain the historical earthquake data set corresponding to the target analysis area and perform preprocessing to obtain the preprocessed historical earthquake data set, where the historical earthquake data set includes area data and earthquake data.

[0057] The historical earthquake data set can be obtained through one or more of the following methods:

[0058] 1. Official earthquake monitoring agencies:

[0059] China Earthquake Administration: The China Earthquake Networks Center provides real-time earthquake data, which can be queried through its official website.

[0060] U.S. Geological Survey (USGS): Provides query services for global earthquake data.

[0061] 2. Data sharing platforms:

[0062] National Earth System Science Data Sharing Service Platform: Provides various geoscience data including earthquake data.

[0063] National Data Sharing and Exchange Platform: Aggregates data resources of various government departments, including earthquake data.

[0064] 3. Academic research institutions:

[0065] Some earthquake laboratories of universities and research institutions will publish the earthquake data sets they have collected or studied.

[0066] 4. International data organizations:

[0067] IRIS (Incorporated Research Institutions for Seismology): Provides global earthquake data.

[0068] EMSC (European-Mediterranean Seismological Centre): Provides seismic data for the European and Mediterranean regions.

[0069] 5. Open-source data projects:

[0070] There are also some open-source projects on GitHub that collect and organize seismic data and can be downloaded and used for free.

[0071] Historical earthquake datasets can include regional data related to regional information and seismic data related to earthquake information. Data related to earthquake information can include origin time, epicenter longitude, epicenter latitude, focal depth, and magnitude type, etc. Regional data related to regional information can include the geological structure, crust type, geomorphic features, climate change, and underground fluid activities of the region.

[0072] Geological structure, crust type, and geomorphic features are all important factors affecting the occurrence of earthquakes. They can cause earthquakes either individually or jointly. The following is a detailed explanation of how these factors affect the occurrence of earthquakes:

[0073] 1. Geological structure:

[0074] Plate movement: The most common cause of earthquakes is the interaction between plates. The mutual extrusion, stretching, or rubbing of plates will accumulate stress in the crust. When the stress exceeds the strength of the rock, fractures will occur, releasing energy and causing earthquakes.

[0075] Fault activity: Faults are fracture zones in the crust. The rocks near active faults will displace under the action of continuous stress, resulting in earthquakes.

[0076] 2. Crust type:

[0077] Thickness and composition of the crust: The thickness and composition of the crust affect its strength and brittleness. Thinner or more brittle crusts are more likely to fracture, thus triggering earthquakes.

[0078] Rock type: Different types of rocks respond differently to stress. For example, some rocks may undergo plastic deformation when stressed and will not fracture immediately; while other rocks may be more brittle and prone to fracture.

[0079] 3. Geomorphic features:

[0080] Mountain formation: The formation of mountains is often related to the collision of plates, and this geological activity may cause earthquakes.

[0081] Geological age: Some geomorphic features, such as the uplift rate of mountains, can indicate the intensity and frequency of crustal activity, thus affecting the occurrence of earthquakes.

[0082] Surface morphology: Surface morphology can reflect the underground geological structure, such as the surface trace of a fault, etc. These areas are usually high-risk zones for earthquakes.

[0083] Generally speaking, geological structure, crust type, and geomorphic features are all potential factors for earthquakes. They determine the frequency and intensity of earthquakes by affecting the accumulation and release of crustal stress. However, the occurrence of an earthquake is a complex natural process, usually the result of the combined action of multiple factors. Therefore, earthquake prediction and risk assessment need to comprehensively consider these factors and other possible factors, such as underground fluid activities, climate change, etc.

[0084] Preprocessing the historical earthquake dataset usually involves correcting or removing abnormal and incorrect acquisition parameters in the earthquake data system. This is because the historical earthquake data obtained is usually historical parameters over several years, decades, or even hundreds of years, and the number of parameters is huge. Therefore, the point-by-point elimination method needs to be used to find obvious abnormal data points, then determine the normal data, and finally perform all possible subset regressions on the remaining data within the abnormal data range. In addition, for possible data missing or anomalies in the dataset, the missing data also needs to be filled.

[0085] The Lagrange interpolation algorithm can be used to fill in the missing data. The specific steps are as follows:

[0086] S21: Assume X = {x0, x1, x2,..., x n} is the interpolation sample set, Y = {y0, y1, y2,..., y n} is the corresponding function value of X, n represents the number of nodes, and x represents the interpolation node;

[0087] S22: Input n + 1 interpolation sample points x0, x1, x2,..., x n and the corresponding function values y0, y1, y2,..., y n ;

[0088] S23: The formula for calculating the nth-order Lagrange basis function is:

[0089]

[0090] where i = 0, 1, 2,...n;

[0091] S24: Calculate the nth-order Lagrange interpolation function:

[0092]

[0093] where l i (x) represents the nth-order Lagrange basis function.

[0094] S25: Input the interpolation node x and substitute it into the interpolation function in step S24 to obtain the calculation result.

[0095] S3: Perform sub-region marking on the preprocessed historical earthquake data set to obtain multiple source sub-regions and multiple non-source sub-regions.

[0096] Match multiple sub-regions according to the preprocessed historical earthquake data set. If it is determined based on the location information of a certain earthquake source in the historical earthquake data set that the earthquake source is located within a certain sub-region, then mark this sub-region as a source sub-region, thereby obtaining one or more source sub-regions, and mark the remaining sub-regions as non-source sub-regions.

[0097] S4: Obtain the data subsets corresponding to the source sub-regions and non-source sub-regions, and input the data subsets into an artificial neural network for feature extraction to obtain the feature vectors of the source sub-regions and the feature vectors of the non-source sub-regions.

[0098] Obtain the source data set corresponding to the source sub-regions and the non-source data set corresponding to the non-source sub-regions in the historical earthquake data set. For the source sub-regions, since an earthquake has occurred in this sub-region, the corresponding data naturally includes regional data related to regional information and earthquake data related to earthquake information. Therefore, set the source data set to include an earthquake data subset and a source region data subset. For the non-source sub-regions, since no earthquake has occurred in this sub-region, the corresponding data naturally only includes regional data related to regional information. Therefore, set the non-source data set to only include a non-source region data subset.

[0099] Input the source region data subset and the earthquake data subset corresponding to the source sub-regions into an artificial neural network for feature extraction respectively to obtain the source region feature vectors Q m (x1, x2, …, x i ) and the earthquake feature vectors P m (y1, y2, …, y j ). Input the non-source region data subset corresponding to the non-source sub-regions into an artificial neural network for feature extraction to obtain the non-source region feature vectors Q n (x1, x2, …, x i ) of each non-source sub-region; where 1 ≤ m ≤ M, and 1 ≤ n ≤ N, M is the number of source sub-regions, N is the number of non-source sub-regions, both M and N are positive integers greater than or equal to 1, i is the number of regional features, j is the number of earthquake features, and both i and j are positive integers greater than or equal to 1.

[0100] S5: Calculate the similarity between the non-seismic source sub-regions and multiple seismic source sub-regions, mark the non-seismic source sub-regions with similarity greater than the preset threshold as potential seismic source sub-regions, and mark the non-seismic source sub-regions with similarity less than or equal to the preset threshold as non-seismic source sub-regions.

[0101] The similarity between sub-regions can be obtained by calculating the similarity between feature vectors. First, calculate the non-seismic source region feature vector Q of any non-seismic source sub-region. n and the seismic source region feature vector Q of any seismic source sub-region. m The similarity between them. Then sum them up to obtain the similarity between the non-seismic source region feature vector Q of any non-seismic source sub-region. n and the seismic source region feature vectors of multiple seismic source sub-regions. sim(Q n ) can be used as the similarity between the non-seismic source sub-region and multiple seismic source sub-regions.

[0102] If sim(Q n ) is greater than the preset threshold, it indicates that the non-seismic source region feature of this non-seismic source sub-region is relatively similar to the seismic source region features of one or more sub-regions in the seismic source sub-regions, and there is also a possibility of earthquake occurrence in this non-seismic source sub-region. Therefore, it is marked as a potential seismic source sub-region. If sim(Q n ) is less than or equal to the preset threshold, it indicates that the non-seismic source region feature of this non-seismic source sub-region is not similar to the seismic source region features of each sub-region in the seismic source sub-regions, and the possibility of earthquake occurrence in this non-seismic source sub-region is extremely low. Therefore, it is marked as a non-seismic source sub-region.

[0103] S6: Use the historical earthquake dataset to train the convolutional neural network to obtain an earthquake prediction model.

[0104] Construct a training sample set from part of the seismic source dataset and non-seismic source dataset according to a preset ratio, use the regional data subset in the remaining seismic source dataset as the test set, and the seismic data subset as the validation set.

[0105] Use the convolutional neural network as the model to be trained, input the training sample set into the convolutional neural network for training. The validation set is used to observe whether the loss function of the convolutional neural network converges during the training process to determine whether to terminate the training, and the test set is used to test the prediction accuracy of the convolutional neural network. Use the stochastic gradient descent method to train the convolutional neural network until the loss function of the neural network converges to the minimum value, and use the trained convolutional neural network as the earthquake prediction model.

[0106] S7: Different monitoring strategies are adopted for the source sub-region, potential source sub-region, and non-seismic source sub-region. According to the monitoring strategies, corresponding regional data subsets are collected from the source sub-region, potential source sub-region, and non-seismic source sub-region, and the regional data subsets are input into the earthquake prediction model for prediction to obtain the earthquake prediction results. An earthquake risk analysis report is output according to the earthquake prediction results.

[0107] The different monitoring strategies can refer to collecting regional data within the sub-region based on different collection frequencies. For example, for the source sub-region, regional data is collected according to the preset collection frequency D1; for the potential source sub-region, regional data is collected according to the preset collection frequency D2; for the non-seismic source sub-region, regional data is collected according to the preset collection frequency D3, where D1 > D2 > D3 > 0. By inputting the collected regional data into the earthquake prediction model to obtain the earthquake prediction results, it is possible to timely detect abnormalities in the regional data in the source sub-region or potential source sub-region, and then determine whether there is a possibility of an earthquake. By outputting an earthquake risk analysis report, it is to prompt the managers within the sub-region to do relevant prevention work and reduce the losses caused by the occurrence of an earthquake.

[0108] In addition, existing monitoring strategies in the prior art can be combined according to actual needs to form multiple monitoring strategies with different priorities, realizing personalized monitoring of different types of regions and improving the fineness of earthquake risk analysis.

[0109] The method for natural earthquake risk analysis based on artificial neural network provided by the embodiment of the present invention realizes the earthquake risk analysis of multiple regions, accurately extracts the characteristics of the earthquake source region through the artificial neural network, thereby judging the potential earthquake regions in the regions where earthquakes have not occurred, and effectively reducing the earthquake risk of the potential earthquake regions; in addition, earthquake prediction is carried out through the convolutional neural network to form an earthquake risk analysis report, providing a scientific basis for managers to make earthquake prevention decisions and further reducing the harm caused by earthquakes.

[0110] Embodiment 2

[0111] As Figure 2 shown, a natural earthquake risk analysis system based on artificial neural network provided by the second embodiment of the present invention is used to implement the method for natural earthquake risk analysis based on artificial neural network described in the above embodiment. The system includes:

[0112] A regional division module, configured to obtain a target analysis region and divide the target analysis region into multiple sub-regions;

[0113] A data processing module, configured to obtain a historical earthquake dataset corresponding to an analysis area and perform preprocessing to obtain a preprocessed historical earthquake dataset, where the historical earthquake dataset includes area data and earthquake data;

[0114] An area marking module, configured to mark sub-areas of the preprocessed historical earthquake dataset to obtain a plurality of earthquake source sub-areas and a plurality of non-earthquake source sub-areas, and further configured to calculate the similarity between the non-earthquake source sub-areas and the plurality of earthquake source sub-areas, mark the non-earthquake source sub-areas with a similarity greater than a preset threshold as potential earthquake source sub-areas, and mark the non-earthquake source sub-areas with a similarity less than or equal to the preset threshold as non-earthquake source sub-areas;

[0115] A feature extraction module, configured to obtain data subsets corresponding to the earthquake source sub-areas and the non-earthquake source sub-areas, input the data subsets into an artificial neural network for feature extraction to obtain feature vectors of the earthquake source sub-areas and feature vectors of the non-earthquake source sub-areas;

[0116] A model training module, configured to train a convolutional neural network based on the historical earthquake dataset to obtain an earthquake prediction model;

[0117] An area monitoring module, configured to adopt different monitoring strategies for the earthquake source sub-areas, potential earthquake source sub-areas and non-earthquake source sub-areas, collect corresponding area data subsets from the earthquake source sub-areas, potential earthquake source sub-areas and non-earthquake source sub-areas according to the monitoring strategies, input the area data subsets into the earthquake prediction model for prediction to obtain an earthquake prediction result, and output an earthquake risk analysis report according to the earthquake prediction result.

[0118] Wherein, the execution process of each module can be executed according to the process steps of a method for analyzing natural earthquake risk based on an artificial neural network in Embodiment 1, and will not be elaborated one by one in this embodiment.

[0119] Embodiment 3

[0120] The third embodiment of the present invention further provides a device for analyzing natural earthquake risk based on an artificial neural network, including a memory and a processor, the memory and the processor are coupled; the memory stores program instructions, and when the program instructions are executed by the processor, the device is enabled to execute the above-mentioned method for analyzing natural earthquake risk based on an artificial neural network.

[0121] Embodiment 4

[0122] The fourth embodiment of the present invention further provides a computer-readable storage medium, including a computer program, when the computer program runs on an electronic device, the electronic device is enabled to execute the above-mentioned method for analyzing natural earthquake risk based on an artificial neural network.

[0123] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) containing computer-usable program code.

[0124] The present application is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram, as well as the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one or more of the flows Figure 1 or blocks or the combination of blocks.

[0125] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing devices to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in Figure 1 one or more of the flows Figure 1 or blocks or the combination of blocks.

[0126] These computer program instructions can also be loaded onto a computer or other programmable data processing devices, such that a series of operation steps are executed on the computer or other programmable devices to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable devices provide steps for implementing the functions specified in Figure 1 one or more of the flows Figure 1 or blocks or the combination of blocks.

[0127] The specific embodiments described above further elaborate on the purpose, technical solutions, and beneficial effects of the present invention. It should be understood that the above descriptions are only specific embodiments of the present invention and are not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for natural earthquake hazard analysis based on artificial neural network, characterized in that, Including: S1: Obtain the target analysis area and divide the target analysis area into multiple sub-areas; S2: Obtain the historical earthquake dataset corresponding to the target analysis area and perform preprocessing to obtain the preprocessed historical earthquake dataset, where the historical earthquake dataset includes regional data and earthquake data; S3: Perform sub-area marking on the preprocessed historical earthquake dataset to obtain multiple source sub-areas and multiple non-source sub-areas; S4: Obtain the data subsets corresponding to the source sub-areas and non-source sub-areas, input the data subsets into an artificial neural network for feature extraction, and obtain the feature vectors of the source sub-areas and the feature vectors of the non-source sub-areas; S5: Calculate the similarity between the non-source sub-areas and multiple source sub-areas, mark the non-source sub-areas with a similarity greater than the preset threshold as potential source sub-areas, and mark the non-source sub-areas with a similarity less than or equal to the preset threshold as non-source sub-areas without earthquakes; S6: Use the historical earthquake dataset to train a convolutional neural network to obtain an earthquake prediction model; S7: Adopt different monitoring strategies for the source sub-areas, potential source sub-areas, and non-source sub-areas without earthquakes, collect the corresponding regional data subsets from the source sub-areas, potential source sub-areas, and non-source sub-areas without earthquakes according to the monitoring strategies, input the regional data subsets into the earthquake prediction model for prediction to obtain earthquake prediction results, and output an earthquake risk analysis report according to the earthquake prediction results.

2. The method for analyzing natural earthquake hazard based on artificial neural network according to claim 1, characterized in that The specific method for obtaining the historical earthquake dataset corresponding to the target analysis area and performing preprocessing includes: Use the point-by-point elimination method to remove abnormal data points in the historical earthquake dataset to obtain the remaining data, and fill in the missing data in the remaining data.

3. The method for analyzing natural earthquake hazards based on an artificial neural network according to claim 2, wherein The method for filling in the missing data in the remaining data uses the Lagrange interpolation algorithm, specifically: S21: Assume that \(X = \{x_0, x_1, x_2, \cdots, x n \}\) is the interpolation sample set, \(Y=\{y_0, y_1, y_2, \cdots, y n \}\) is the corresponding function values of \(X\), \(n\) represents the number of nodes, and \(x\) represents the interpolation node; S22: Input n + 1 interpolation sample points x0, x1, x2, ..., x n and the corresponding function values y0, y1, y2, ..., y n ; S23: Calculate the nth-order Lagrange basis function: where i = 0, 1, 2,... n; S24: Calculate the nth-order Lagrange interpolation function: Among them, l i (x) represents the Lagrange basis function of order n; S25: Input the interpolation point x and substitute it into the interpolation function in step S24 to obtain the calculation result.

4. The method for analyzing natural earthquake hazard based on artificial neural network according to claim 1, characterized in that The specific method for performing sub-area marking on the preprocessed historical earthquake dataset includes: Judge whether the location information of the earthquake source in the historical earthquake dataset is within a certain sub-area. If so, mark the sub-area as a source sub-area; if not, mark the sub-area as a non-source sub-area.

5. The natural earthquake hazard analysis method based on artificial neural network according to claim 1, characterized in that The method for obtaining the data subsets corresponding to the source sub-areas and non-source sub-areas, and inputting the data subsets into an artificial neural network for feature extraction specifically includes: Obtain the source dataset corresponding to the source sub-areas and the non-source dataset corresponding to the non-source sub-areas from the historical earthquake dataset. The source dataset includes an earthquake data subset and a source area data subset, and the non-source dataset includes a non-source area data subset; Input the earthquake data subset and the source area data subset into an artificial neural network for feature extraction respectively to obtain the earthquake feature vectors and source area feature vectors corresponding to each source sub-area; Input the non-source area data subset into an artificial neural network for feature extraction to obtain the non-source area feature vectors corresponding to each non-source sub-area.

6. The method for analyzing natural earthquake hazards based on an artificial neural network according to claim 5, characterized in that, Training a convolutional neural network using a historical earthquake dataset to obtain an earthquake prediction model, specifically including: Constructing a training sample set from the earthquake source dataset and the non-earthquake source dataset according to a preset ratio, using the regional data subset in the remaining earthquake source dataset as the test set, and the earthquake data subset as the validation set; Using a convolutional neural network as the model to be trained, inputting the training sample set into the convolutional neural network for training. The validation set is used to observe whether the loss function of the convolutional neural network converges during training to determine whether to terminate training, and the test set is used to test the prediction accuracy of the convolutional neural network. The convolutional neural network is trained using the stochastic gradient descent method until the loss function of the neural network converges to the minimum value, and the trained convolutional neural network is used as the earthquake prediction model.

7. The method for analyzing natural earthquake hazard based on artificial neural network according to claim 1, characterized in that, The monitoring strategy is to collect regional data in sub-regions at different acquisition frequencies. The acquisition frequency of the earthquake source sub-region is greater than that of the potential earthquake source sub-region, and the acquisition frequency of the potential earthquake source sub-region is greater than that of the non-earthquake source sub-region.

8. A natural earthquake hazard analysis system based on an artificial neural network, which is used to implement the natural earthquake hazard analysis method based on an artificial neural network as described in claim 1, characterized in that, Including: A regional division module for obtaining a target analysis region and dividing the target analysis region into multiple sub-regions; A data processing module for obtaining the historical earthquake dataset corresponding to the analysis region and performing preprocessing to obtain the preprocessed historical earthquake dataset, where the historical earthquake dataset includes regional data and earthquake data; A regional marking module for marking the preprocessed historical earthquake dataset with sub-regions to obtain multiple earthquake source sub-regions and multiple non-earthquake source sub-regions, and also for calculating the similarity between the non-earthquake source sub-regions and the multiple earthquake source sub-regions, marking the non-earthquake source sub-regions with a similarity greater than a preset threshold as potential earthquake source sub-regions, and marking the non-earthquake source sub-regions with a similarity less than or equal to the preset threshold as non-earthquake source sub-regions; A feature extraction module for obtaining the data subsets corresponding to the earthquake source sub-regions and the non-earthquake source sub-regions, inputting the data subsets into an artificial neural network for feature extraction to obtain the feature vectors of the earthquake source sub-regions and the feature vectors of the non-earthquake source sub-regions; A model training module for training a convolutional neural network based on the historical earthquake dataset to obtain an earthquake prediction model; A regional monitoring module for adopting different monitoring strategies for the earthquake source sub-regions, potential earthquake source sub-regions, and non-earthquake source sub-regions, collecting the corresponding regional data subsets from the earthquake source sub-regions, potential earthquake source sub-regions, and non-earthquake source sub-regions according to the monitoring strategy, inputting the regional data subsets into the earthquake prediction model for prediction to obtain earthquake prediction results, and outputting an earthquake hazard analysis report according to the earthquake prediction results.

9. A natural earthquake hazard analysis device based on an artificial neural network, characterized in that, The device includes: a memory and a processor, the memory and the processor are coupled; the memory stores program instructions, and when the program instructions are executed by the processor, the device executes the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Including a computer program, when the computer program runs on an electronic device, the electronic device executes the method according to any one of claims 1 to 7.

Images