Intelligent geological disaster risk assessment method and system

Abstract

The invention discloses an intelligent geological disaster risk assessment method and system. The method comprises the steps of data acquisition, data preprocessing, geological disaster risk assessment model construction, parameter search and risk assessment. The invention relates to the technical field of disaster risk assessment, in particular to an intelligent geological disaster risk assessment method and system.According to the scheme, the dynamic adjustment ability of data distribution and the expression ability of time dimension data are enhanced through normalization enhancement, time weight definition, time fusion, feature mapping function design and feature transformation; the capturing capability of the model on complex features is improved; by designing a kernel function, designing a manifold regularization item, designing a parameter updating rule and introducing topological constraints, the convergence speed, the calculation efficiency and the expression ability of the model are improved; by setting the search step length, designing the search guide factor and selecting the dynamic path, the systematicness and scientificity of the search are improved, and the efficiency and precision of the search are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of disaster risk assessment, and in particular to an intelligent geological disaster risk assessment method and system. Background Art

[0002] An intelligent geological disaster risk assessment method and system is a comprehensive solution that uses artificial intelligence technology to efficiently and accurately assess geological disaster risks. It uses artificial intelligence algorithms to build an intelligent risk assessment model, providing a scientific basis for geological disaster early warning, prevention and control, and emergency management, and realizing a comprehensive analysis and prediction of geological disaster risks.

[0003] Traditional geological hazard risk assessment methods have problems such as insufficient data utilization, limited feature extraction capabilities, and insufficient consideration of time dynamics; traditional geological hazard risk assessment models have problems such as insufficient capture of local geometric structure of data, single regularization method, low model convergence speed and computational efficiency; traditional parameter search methods have problems such as low search efficiency, lack of dynamic adjustment and direction guidance capabilities in parameter selection, and lack of dynamics and adaptability in path selection. Summary of the Invention

[0004] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides an intelligent geological hazard risk assessment method and system. In view of the problems of insufficient data utilization, limited feature extraction capability and insufficient consideration of time dynamics in traditional geological hazard risk assessment methods, this solution enhances the dynamic adjustment capability of data distribution by enhancing normalization, defining time weights, time fusion, designing feature mapping functions and feature transformations, better captures key information in time series data, improves the modeling capability of time dynamics, enhances the expression capability of time dimension data, and improves the model's ability to capture complex features; in view of the problems of insufficient capture of local geometric structure of data, single regularization method, low convergence speed and computational efficiency of traditional geological hazard risk assessment models, this solution designs a kernel function The method adopts the multi-dimensional space-separated list to optimize the geological disaster risk assessment model, and designs the manifold regularization term, designs the parameter update rules and introduces topological constraints to better capture the local geometric structure and global topological characteristics of the data, and optimizes more efficiently in the manifold space, thereby improving the convergence speed and computational efficiency of the model, and enhancing the expressiveness, stability and generalization performance of the geological disaster risk assessment model. In view of the problems of low search efficiency, lack of dynamic adjustment and direction guidance ability of parameter selection, and lack of dynamic and adaptability of path selection in traditional parameter search methods, this scheme improves the systematicness, scientificity and coverage of the search by building a search framework, generating initial parameter search points, setting search step size, designing search guidance factors, dynamic path selection and iterative exploration, so that the search process can be flexibly adjusted according to the current state, thereby improving the efficiency, accuracy and intelligence of the search.

[0005] The technical solution adopted by the present invention is as follows: an intelligent geological disaster risk assessment method, the method comprising the following steps:

[0006] Step S1: data collection;

[0007] Step S2: data preprocessing;

[0008] Step S3: constructing a geological hazard risk assessment model;

[0009] Step S4: parameter search;

[0010] Step S5: Risk assessment.

[0011] Furthermore, in step S1, the data collection is to collect topographic data, geological data, meteorological data, hydrological data and historical disaster data; the topographic data includes slope, slope aspect and land type; the geological data includes stratum distribution, lithology and soil type; the meteorological data includes rainfall, temperature, humidity, wind speed and wind direction; the hydrological data includes river flow, groundwater level, flow velocity and flow direction; the historical disaster data includes disaster type, occurrence time, occurrence scale, caused losses and frequency of disaster occurrence.

[0012] Furthermore, in step S2, the data preprocessing specifically includes the following steps:

[0013] Step S21: Setting label data, dividing the collected data into three parts according to the frequency of disasters in the historical disaster data of the collected data, setting the 30% of data with the highest frequency of disasters as high geological disaster risk data; setting the 30% of data with the lowest frequency of disasters as low geological disaster risk data; and setting the remaining data as medium geological disaster risk data; creating a data feature: geological disaster risk level, including three risk levels of high risk, medium risk and low risk, and the three data sets corresponding to the three risk levels; setting the geological disaster risk level as label data;

[0014] Step S22: Enhanced normalization, expressed as follows:

[0015] ;

[0016] in, Represents the data after enhanced normalization, Represents the original data, represents the mean of the original data, represents the standard deviation of the original data, e represents the natural constant, represents the expansion coefficient, and represents the sensitivity parameter;

[0017] Step S23: define the time weight, which is expressed as follows:

[0018] ;

[0019] Where m represents the index of the time point, represents the time weight of the mth time point, represents the total number of time points, The index of the time point. represents the control coefficient, represents the original data at the mth time point, represents the entropy of the original data at the mth time point, Indicates the The original data at each time point, Indicates the The entropy of the raw data at each time point;

[0020] Step S24: time fusion, which is expressed as follows:

[0021] ;

[0022] in, Represents the data after time fusion, represents the hyperbolic tangent function, represents the mean of the original data at the mth time point, represents the standard deviation of the original data at the mth time point, represents taking the logarithm, Indicates taking the absolute value;

[0023] Step S25: Design a feature mapping function, which is expressed as follows:

[0024] ;

[0025] in, Represents the input value of the feature mapping function, represents the feature mapping function;

[0026] Step S26: Feature transformation, constructing a neural network to perform feature transformation, which is expressed as follows:

[0027] ;

[0028] in, represents the feature transformation function, represents the linear rectification function, 、 and represents the feature transformation weight, 、 and represents the feature transformation bias, Represents the data after feature transformation, represents the mean of the data after time fusion, represents the variance of the data after time fusion, Indicates the modulus length.

[0029] Furthermore, in step S3, the construction of the geological disaster risk assessment model specifically includes the following steps:

[0030] Step S31: Design a kernel function, which is expressed as follows:

[0031] ;

[0032] Among them, i and j represent the index of the data used for model training, and Indicates the i-th and j-th data, Represents the manifold kernel function value between data i and data j, M represents the manifold space, represents the shortest path distance between data i and data j on the manifold, represents the standard deviation of the data used for model training, k represents the index of the local coordinate function, D represents the total number of local coordinate functions, represents the local coordinate weight, represents the local coordinate function value of data i on the manifold, express The gradient, represents the local coordinate function value of data j on the manifold, express The gradient, represents a local regulatory factor;

[0033] Step S32: Design a manifold regularization term, which is expressed as follows:

[0034] ;

[0035] in, represents the manifold weight parameter, represents the regularization term of the manifold weight parameter, represents the integration on the manifold space, represents the gradient operator on the manifold, represents data on the manifold, represents the manifold bias parameter;

[0036] Step S33: Determine the initial optimization target, which is expressed as follows:

[0037] ;

[0038] in, represents the relaxation factor of the i-th data, The goal is to find the parameters 、 and Minimize the value of the target expression, C represents the trade-off parameter, and N represents the total number of data used for model training. and represents the regularization coefficient, Represents the Hessian matrix of the classification hyperplane of the i-th data;

[0039] Step S34: Design parameter update rules, as shown below:

[0040] ;

[0041] in, Indicates the number of times the model parameters are updated, Indicates the The model parameters when the model parameters are updated, represents the exponential mapping function on the manifold, represents the manifold parameter update rate, represents the objective function, Indicates the The model parameters when the model parameters are updated, Represents the objective function on the manifold with respect to the parameter gradient;

[0042] Step S35: introduce topological constraints, which are expressed as follows:

[0043] ;

[0044] in, represents the topological constraint term introduced on the manifold weight parameter, represents the subset of the manifold on the kth local coordinate function, represents the integration over the space of manifold subsets;

[0045] Step S36: Determine the final optimization target, which is expressed as follows:

[0046] ;

[0047] in, The weight coefficient representing the topological constraint.

[0048] Furthermore, in step S4, the parameter search specifically includes the following steps:

[0049] Step S41: Build a search framework, determine search parameters, including local adjustment factors, trade-off parameters, and regularization coefficients, and set the accuracy of the model as the fitness value of the parameter point;

[0050] Step S42: Generate an initial parameter search point, which is expressed as follows:

[0051] ;

[0052] Where h represents the index of the parameter search point, represents the generated h-th initial parameter search point, represents the maximum value of the search space, represents the minimum value of the search space, Indicates a A random number uniformly distributed on

[0053] Step S43: Set the search step size, as shown below:

[0054] ;

[0055] Among them, t represents the number of parameter explorations, Indicates the search step size of the hth parameter search point during the t+1th parameter search, Indicates the search step size of the hth parameter search point during the tth parameter search, represents the search rate, Indicates the fitness value of the hth parameter search point during the tth parameter search, represents the highest fitness value, represents the lowest fitness value, represents the average position of the initial parameter search points;

[0056] Step S44: Design a search guide factor, which is expressed as follows:

[0057] ;

[0058] Among them, q represents the index of the parameter search direction, Indicates the search guidance factor of the qth search direction when the parameter search point performs the t+1th parameter search, represents the decay rate, Indicates the search guidance factor of the qth search direction when the parameter search point performs the tth parameter search, Indicates the parameter position found in the qth search direction when the parameter search point performs the tth parameter search. express The fitness value of represents the initial guidance strength;

[0059] Step S45: Dynamic path selection, as shown below:

[0060] ;

[0061] in, It represents the probability that the parameter search point chooses the path of the qth search direction when performing the t+1th parameter search. Indicates the fitness change value of the parameter search point in the qth search direction when performing the t+1th parameter search. Indicates the maximum search guidance factor of the parameter search point when performing the t+1th parameter search, Indicates the maximum fitness change of the parameter search point during the t+1th parameter search;

[0062] Step S46: Iterative exploration, setting the fitness threshold, each time an iterative search is performed, the probability of all parameter search points selecting each search direction is calculated respectively, the direction with the highest probability is selected, and the distance of one search step is advanced, and the fitness value of the parameter search point at this time is calculated; when the fitness value is greater than the fitness threshold, the search ends, and the parameter with the highest fitness value at this time is set as the model parameter; otherwise, the number of iterations is increased by one, and the next iterative search is performed.

[0063] Furthermore, in step S5, the risk assessment is to collect topographic data, geological data, meteorological data and hydrological data of the assessment area, input them into a geological disaster risk assessment model, and the model outputs the geological disaster risk level of the assessment area.

[0064] The present invention provides an intelligent geological disaster risk assessment system, which includes a data acquisition module, a data preprocessing module, a geological disaster risk assessment model construction module, a parameter search module and a risk assessment module;

[0065] The data acquisition module collects topographic data, geological data, meteorological data, hydrological data and historical disaster data, and sends the data to the data preprocessing module;

[0066] The data preprocessing module receives the data sent by the data acquisition module, preprocesses the data, and sends the data to the geological disaster risk assessment model construction module;

[0067] The geological disaster risk assessment model construction module receives the data sent by the data preprocessing module, constructs the geological disaster risk assessment model module, and sends the data to the parameter search module and the risk assessment module;

[0068] The parameter search module receives data sent by the geological disaster risk assessment model construction module, searches for parameters of the model, and sends the data to the risk assessment module;

[0069] The risk assessment module receives data sent by the geological disaster risk assessment model construction module and the parameter search module, collects terrain data, geological data, meteorological data and hydrological data of the assessment area, inputs the geological disaster risk assessment model, and the model outputs the geological disaster risk level of the assessment area.

[0070] The beneficial effects achieved by the present invention using the above scheme are as follows:

[0071] (1) In view of the problems of insufficient data utilization, limited feature extraction capability and insufficient consideration of time dynamics in traditional geological hazard risk assessment methods, this scheme enhances the dynamic adjustment capability of data distribution by enhancing normalization, defining time weights, time fusion, designing feature mapping functions and feature transformations, better capturing key information in time series data, improving the modeling capability of time dynamics, enhancing the expression capability of time dimension data, and improving the model's ability to capture complex features.

[0072] (2) In view of the problems of insufficient capture of local geometric structure of data, single regularization method, low convergence speed and computational efficiency of traditional geological hazard risk assessment models, this scheme designs kernel functions, manifold regularization terms, parameter update rules and introduces topological constraints to better capture the local geometric structure and global topological characteristics of data, optimize more efficiently in manifold space, improve the convergence speed and computational efficiency of the model, and enhance the expressive power, stability and generalization performance of the geological hazard risk assessment model.

[0073] (3) In response to the problems of low search efficiency, lack of dynamic adjustment and direction guidance capabilities in parameter selection, and lack of dynamic and adaptable path selection in traditional parameter search methods, this solution improves the systematicity, scientificity and coverage of the search by building a search framework, generating initial parameter search points, setting search steps, designing search guidance factors, dynamic path selection and iterative exploration, so that the search process can be flexibly adjusted according to the current status, thereby improving the efficiency, accuracy and intelligence level of the search. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 A schematic diagram of an intelligent geological hazard risk assessment method provided by the present invention;

[0075] Figure 2 A schematic diagram of an intelligent geological disaster risk assessment system provided by the present invention;

[0076] Figure 3 Schematic diagram of data preprocessing;

[0077] Figure 4 Schematic diagram for constructing a geological hazard risk assessment model;

[0078] Figure 5 Schematic diagram of parameter search.

[0079] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0080] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0081] In the description of the present invention, it should be understood that terms such as "upper", "lower", "front", "back", "left", "right", "top", "bottom", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction. Therefore, they should not be understood as limiting the present invention.

[0082] Example 1, see Figure 1 The present invention provides an intelligent geological disaster risk assessment method, which includes the following steps:

[0083] Step S1: Data collection, collecting topographic data, geological data, meteorological data, hydrological data and historical disaster data;

[0084] Step S2: Data preprocessing, which involves setting label data, enhancing normalization, defining time weights, time fusion, designing feature mapping functions, and feature transformations.

[0085] Step S3: constructing a geological hazard risk assessment model by designing a kernel function, designing a manifold regularization term, determining an initial optimization target, designing a parameter update rule, introducing a topological constraint, and determining a final optimization target;

[0086] Step S4: Parameter search, which involves building a search framework, generating initial parameter search points, setting search step lengths, designing search guide factors, dynamic path selection, and iterative exploration to perform parameter search.

[0087] Step S5: Risk assessment, collecting topographic data, geological data, meteorological data and hydrological data of the assessment area, inputting them into the geological disaster risk assessment model, and the model outputs the geological disaster risk level of the assessment area.

[0088] Example 2, see Figure 1 and Figure 3 This embodiment is based on the above embodiment, and the data preprocessing specifically includes the following steps:

[0089] Step S21: Setting label data, dividing the collected data into three parts according to the frequency of disasters in the historical disaster data of the collected data, setting the 30% of data with the highest frequency of disasters as high geological disaster risk data; setting the 30% of data with the lowest frequency of disasters as low geological disaster risk data; and setting the remaining data as medium geological disaster risk data; creating a data feature: geological disaster risk level, including three risk levels of high risk, medium risk and low risk, and the three data sets corresponding to the three risk levels; setting the geological disaster risk level as label data;

[0090] Step S22: Enhanced normalization, expressed as follows:

[0091] ;

[0092] in, Represents the data after enhanced normalization, Represents the original data, represents the mean of the original data, represents the standard deviation of the original data, e represents the natural constant, represents the expansion coefficient, and represents the sensitivity parameter;

[0093] Step S23: define the time weight, which is expressed as follows:

[0094] ;

[0095] Where m represents the index of the time point, represents the time weight of the mth time point, represents the total number of time points, The index of the time point. represents the control coefficient, represents the original data at the mth time point, represents the entropy of the original data at the mth time point, Indicates the The original data at each time point, Indicates the The entropy of the raw data at each time point;

[0096] Step S24: Time fusion, fusing the data at multiple time points, as shown below:

[0097] ;

[0098] in, Represents the data after time fusion, represents the hyperbolic tangent function, represents the mean of the original data at the mth time point, represents the standard deviation of the original data at the mth time point, represents taking the logarithm, Indicates taking the absolute value;

[0099] Step S25: Design a feature mapping function and introduce the idea of ​​segmented mapping, which is expressed as follows:

[0100] ;

[0101] in, Represents the input value of the feature mapping function, represents the feature mapping function;

[0102] Step S26: Feature transformation, constructing a neural network to perform feature transformation, which is expressed as follows:

[0103] ;

[0104] in, represents the feature transformation function, represents the linear rectification function, 、 and represents the feature transformation weight, 、 and represents the feature transformation bias, Represents the data after feature transformation, represents the mean of the data after time fusion, represents the variance of the data after time fusion, Indicates the modulus length.

[0105] By performing the above operations, in order to address the problems of insufficient data utilization, limited feature extraction capabilities, and insufficient consideration of time dynamics in traditional geological hazard risk assessment methods, this solution enhances the dynamic adjustment capability of data distribution by enhancing normalization, defining time weights, time fusion, designing feature mapping functions and feature transformations, better capturing key information in time series data, improving the modeling capability of time dynamics, enhancing the expression capability of time dimension data, and improving the model's ability to capture complex features.

[0106] Example 3, see Figure 1 and Figure 4This embodiment is based on the above embodiment, and the construction of the geological disaster risk assessment model specifically includes the following steps:

[0107] Step S31: Design a kernel function, consider the local geometric structure of the data, and introduce a manifold kernel function, which is expressed as follows:

[0108] ;

[0109] Among them, i and j represent the index of the data used for model training, and Indicates the i-th and j-th data, Represents the manifold kernel function value between data i and data j, M represents the manifold space, represents the shortest path distance between data i and data j on the manifold, represents the standard deviation of the data used for model training, k represents the index of the local coordinate function, D represents the total number of local coordinate functions, represents the local coordinate weight, represents the local coordinate function value of data i on the manifold, express The gradient, represents the local coordinate function value of data j on the manifold, express The gradient, represents a local regulatory factor;

[0110] Step S32: Design a manifold regularization term to penalize the curvature change of the classification hyperplane on the manifold, which is expressed as follows:

[0111] ;

[0112] in, represents the manifold weight parameter, represents the regularization term of the manifold weight parameter, represents the integration on the manifold space, represents the gradient operator on the manifold, represents data on the manifold, represents the manifold bias parameter;

[0113] Step S33: Determine the initial optimization target, which is expressed as follows:

[0114] ;

[0115] in, represents the relaxation factor of the i-th data, The goal is to find the parameters 、 and Minimize the value of the target expression, C represents the trade-off parameter, and N represents the total number of data used for model training. and represents the regularization coefficient, Represents the Hessian matrix of the classification hyperplane of the i-th data;

[0116] Step S34: Design parameter update rules and use geodesic gradient descent to update model parameters, which is expressed as follows:

[0117] ;

[0118] in, Indicates the number of times the model parameters are updated, Indicates the The model parameters when the model parameters are updated, represents the exponential mapping function on the manifold, represents the manifold parameter update rate, represents the objective function, Indicates the The model parameters when the model parameters are updated, Represents the objective function on the manifold with respect to the parameter gradient;

[0119] Step S35: introduce topological constraints, which are expressed as follows:

[0120] ;

[0121] in, represents the topological constraint term introduced on the manifold weight parameter, represents the subset of the manifold on the kth local coordinate function, represents the integration over the space of manifold subsets;

[0122] Step S36: Determine the final optimization target, which is expressed as follows:

[0123] ;

[0124] in, The weight coefficient representing the topological constraint.

[0125] By performing the above operations, this scheme addresses the problems of traditional geological hazard risk assessment models, such as insufficient capture of the local geometric structure of the data, single regularization method, low model convergence speed and computational efficiency. By designing kernel functions, manifold regularization terms, parameter update rules and introducing topological constraints, it can better capture the local geometric structure and global topological characteristics of the data, optimize more efficiently in the manifold space, improve the convergence speed and computational efficiency of the model, and enhance the expressiveness, stability and generalization performance of the geological hazard risk assessment model.

[0126] Example 4, see Figure 1 and Figure 5 This embodiment is based on the above embodiment, and the parameter search specifically includes the following steps:

[0127] Step S41: Build a search framework, determine search parameters, including local adjustment factors, trade-off parameters, and regularization coefficients, and set the accuracy of the model as the fitness value of the parameter point;

[0128] Step S42: Generate an initial parameter search point, which is expressed as follows:

[0129] ;

[0130] Where h represents the index of the parameter search point, represents the generated h-th initial parameter search point, represents the maximum value of the search space, represents the minimum value of the search space, Indicates a A random number uniformly distributed on

[0131] Step S43: Set the search step size, and dynamically adjust the moving step size of the parameter search point by combining the current fitness value with the global optimal and worst fitness values, as shown below:

[0132] ;

[0133] Among them, t represents the number of parameter explorations, Indicates the search step size of the hth parameter search point during the t+1th parameter search, Indicates the search step size of the hth parameter search point during the tth parameter search, represents the search rate, Indicates the fitness value of the hth parameter search point during the tth parameter search, represents the highest fitness value, represents the lowest fitness value, represents the average position of the initial parameter search points;

[0134] Step S44: Design a search guide factor, which is expressed as follows:

[0135] ;

[0136] Among them, q represents the index of the parameter search direction, Indicates the search guidance factor of the qth search direction when the parameter search point performs the t+1th parameter search, represents the decay rate, Indicates the search guidance factor of the qth search direction when the parameter search point performs the tth parameter search, Indicates the parameter position found in the qth search direction when the parameter search point performs the tth parameter search. express The fitness value of represents the initial guidance strength;

[0137] Step S45: Dynamic path selection, combining the search guidance factor and the fitness change gradient, dynamically calculates the probability of selecting each search direction, which is expressed as follows:

[0138] ;

[0139] in, It represents the probability that the parameter search point chooses the path of the qth search direction when performing the t+1th parameter search. Indicates the fitness change value of the parameter search point in the qth search direction when performing the t+1th parameter search. Indicates the maximum search guidance factor of the parameter search point when performing the t+1th parameter search, Indicates the maximum fitness change of the parameter search point during the t+1th parameter search;

[0140] Step S46: Iterative exploration, setting the fitness threshold, each time an iterative search is performed, the probability of all parameter search points selecting each search direction is calculated respectively, the direction with the highest probability is selected, and the distance of one search step is advanced, and the fitness value of the parameter search point at this time is calculated; when the fitness value is greater than the fitness threshold, the search ends, and the parameter with the highest fitness value at this time is set as the model parameter; otherwise, the number of iterations is increased by one, and the next iterative search is performed.

[0141] By performing the above operations, the problems of low search efficiency, lack of dynamic adjustment and direction guidance capabilities in parameter selection, and lack of dynamic and adaptable path selection in traditional parameter search methods have been addressed. This solution improves the systematicity, scientificity and coverage of the search by building a search framework, generating initial parameter search points, setting search steps, designing search guidance factors, dynamic path selection and iterative exploration, so that the search process can be flexibly adjusted according to the current status, thereby improving the efficiency, accuracy and intelligence of the search.

[0142] Example 5, see Figure 1 and Figure 2 , this embodiment is based on the above embodiment, and the present invention provides an intelligent geological disaster risk assessment system, including a data acquisition module, a data preprocessing module, a geological disaster risk assessment model construction module, a parameter search module and a risk assessment module;

[0143] The data acquisition module collects topographic data, geological data, meteorological data, hydrological data and historical disaster data, and sends the data to the data preprocessing module;

[0144] The data preprocessing module receives the data sent by the data acquisition module, preprocesses the data, and sends the data to the geological disaster risk assessment model construction module;

[0145] The geological disaster risk assessment model construction module receives the data sent by the data preprocessing module, constructs the geological disaster risk assessment model module, and sends the data to the parameter search module and the risk assessment module;

[0146] The parameter search module receives data sent by the geological disaster risk assessment model construction module, searches for parameters of the model, and sends the data to the risk assessment module;

[0147] The risk assessment module receives data sent by the geological disaster risk assessment model construction module and the parameter search module, collects terrain data, geological data, meteorological data and hydrological data of the assessment area, inputs the geological disaster risk assessment model, and the model outputs the geological disaster risk level of the assessment area.

[0148] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0149] While the embodiments of the present invention have been shown and described, it will be apparent to those skilled in the art that various changes, modifications, substitutions, and alterations can be made to the embodiments without departing from the principles and spirit of the invention.

[0150] The present invention and its embodiments are described above. This description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if a person skilled in the art is inspired by this and, without departing from the purpose of the present invention, designs structures and embodiments similar to this technical solution without inventiveness, they shall fall within the scope of protection of the present invention.

Claims

1. An intelligent geological disaster risk assessment method, characterized in that: The method comprises the following steps: Step S1: Data collection, collecting topographic data, geological data, meteorological data, hydrological data and historical disaster data; Step S2: Data preprocessing, which is performed by setting label data, enhancing normalization, defining time weights, time fusion, designing feature mapping functions and feature transformations; Step S3: constructing a geological hazard risk assessment model by designing a kernel function, designing a manifold regularization term, determining an initial optimization target, designing a parameter update rule, introducing a topological constraint, and determining a final optimization target; Step S31: design a kernel function, which is expressed as follows: ； Where i and j represent the index of the data used for model training, e represents a natural constant, and represents the i-th and j-th data, represents the manifold kernel function value between data i and data j, M represents the manifold space, represents the shortest path distance between data i and data j on the manifold, represents the standard deviation of the data used for model training, k represents the index of the local coordinate function, D represents the total number of local coordinate functions, represents the local coordinate weight, represents the local coordinate function value of data i on the manifold, express The gradient of represents the local coordinate function value of data j on the manifold, express The gradient of represents a local regulatory factor; Step S4: parameter search, which is performed by building a search framework, generating initial parameter search points, setting search step lengths, designing search guide factors, dynamic path selection, and iterative exploration; Step S5: Risk assessment, collecting topographic data, geological data, meteorological data and hydrological data of the assessment area, inputting into the geological disaster risk assessment model, and the model outputs the geological disaster risk level of the assessment area.

2. An intelligent geological disaster risk assessment method according to claim 1, characterized in that: In step S3, the construction of the geological disaster risk assessment model specifically includes the following steps: Step S31: designing a kernel function; Step S32: Design a manifold regularization term, which is expressed as follows: ； in, represents the manifold weight parameter, represents the regularization term of the manifold weight parameter, represents the integration on the manifold space, Indicates the modulus length, represents the gradient operator on the manifold, represents data on the manifold, represents the manifold bias parameter; Step S33: Determine the initial optimization target, which is expressed as follows: ； in, represents the relaxation factor of the i-th data, The goal is to find the parameter , and Make the value of the objective expression minimum, C represents the trade-off parameter, N represents the total number of data used for model training, and represents the regularization coefficient, The Hessian matrix representing the classification hyperplane of the i-th data; Step S34: Design parameter update rules, as shown below: ； in, Indicates the number of model parameter updates, Indicates The model parameters at the time of model parameter update, represents the exponential mapping function on the manifold, represents the manifold parameter update rate, represents the objective function, Indicates The model parameters at the time of model parameter update, Represents the objective function on the manifold with respect to the parameter The gradient of Step S35: introduce topological constraints, which are expressed as follows: ； in, represents the topological constraint term introduced by the manifold weight parameter, represents the manifold subset on the kth local coordinate function, represents the integration over the space of manifold subsets; Step S36: Determine the final optimization target, which is expressed as follows: ； in, Represents the weight coefficient of the topological constraint.

3. The intelligent geological disaster risk assessment method according to claim 1 is characterized in that: In step S2, the data preprocessing specifically includes the following steps: Step S21: setting label data, dividing the collected data into three parts according to the frequency of disasters in the historical disaster data of the collected data, setting the 30% data with the highest frequency of disasters as high geological disaster risk data; setting the 30% data with the lowest frequency of disasters as low geological disaster risk data; and setting the remaining data as medium geological disaster risk data; creating a data feature: geological disaster risk level, including three risk levels of high risk, medium risk and low risk, and the three parts of data correspond to three risk levels; setting the geological disaster risk level as label data; Step S22: Enhanced normalization, expressed as follows: ； in, Represents the data after enhanced normalization. Represents the original data, represents the mean of the original data, represents the standard deviation of the original data, represents the expansion factor, and represents the sensitivity parameter; Step S23: define the time weight, which is expressed as follows: ； Where m represents the index of the time point, represents the time weight of the mth time point, represents the total number of time points, The index of the time point. represents the control coefficient, represents the original data at the mth time point, represents the entropy of the original data at the mth time point, Indicates The original data at each time point, Indicates The entropy of the raw data at each time point; Step S24: time fusion, expressed as follows: ； in, Represents the data after time fusion, represents the hyperbolic tangent function, represents the mean of the original data at the mth time point, represents the standard deviation of the original data at the mth time point, represents taking the logarithm, Indicates taking the absolute value; Step S25: Design a feature mapping function, which is expressed as follows: ； in, represents the input value of the feature mapping function, represents the feature mapping function; Step S26: feature transformation, constructing a neural network to perform feature transformation, which is expressed as follows: ； in, represents the feature transformation function, represents the linear rectification function, , and represents the feature transformation weight, , and represents the feature transformation bias, Represents the data after feature transformation, represents the mean of the data after time fusion, Represents the variance of the data after time fusion.

4. The intelligent geological disaster risk assessment method according to claim 1 is characterized in that: In step S4, the parameter search specifically includes the following steps: Step S41: build a search framework, determine search parameters, including local adjustment factors, trade-off parameters and regularization coefficients, and set the accuracy of the model as the fitness value of the parameter point; Step S42: Generate an initial parameter search point, as shown below: ； Where h represents the index of the parameter search point, represents the generated h-th initial parameter search point, represents the maximum value of the search space, represents the minimum value of the search space, Represents a A random number uniformly distributed on Step S43: Set the search step length, as shown below: ； Among them, t represents the number of parameter explorations, Indicates the search step length of the hth parameter search point during the t+1th parameter search, Indicates the search step length of the hth parameter search point during the tth parameter search, represents the search rate, Indicates the fitness value of the hth parameter search point during the tth parameter search, represents the highest fitness value, represents the lowest fitness value, represents the average position of the initial parameter search points; Step S44: Design a search guide factor, which is expressed as follows: ； Among them, q represents the index of the parameter search direction, It represents the search guidance factor of the qth search direction when the parameter search point performs the t+1th parameter search. represents the decay rate, It represents the search guidance factor of the qth search direction when the parameter search point performs the tth parameter search. Indicates the parameter position found in the qth search direction when the parameter search point performs the tth parameter search. express The fitness value of represents the initial guidance strength; Step S45: Dynamic path selection, represented as follows: ； in, represents the probability that the parameter search point chooses the path in the qth search direction when performing the t+1th parameter search. It indicates the fitness change value of the parameter search point in the qth search direction when performing the t+1th parameter search. It represents the maximum search guide factor of the parameter search point when performing the t+1th parameter search. Indicates the maximum fitness change of the parameter search point during the t+1th parameter search; Step S46: Iterative exploration, setting the fitness threshold, each time an iterative search is performed, the probability of all parameter search points selecting each search direction is calculated respectively, the direction with the highest probability is selected, and the distance of one search step is advanced to calculate the fitness value of the parameter search point at this time; when the fitness value is greater than the fitness threshold, the search ends, and the parameter with the highest fitness value at this time is set as the model parameter; otherwise, the number of iterations is increased by one, and the next iterative search is performed.

5. The intelligent geological disaster risk assessment method according to claim 1 is characterized in that: In step S1, the data collection is to collect terrain data, geological data, meteorological data, hydrological data and historical disaster data; the terrain data includes slope, slope direction and land type; the geological data includes stratum distribution, lithology and soil type; the meteorological data includes rainfall, temperature, humidity, wind speed and wind direction; the hydrological data includes river flow, groundwater level, flow rate and flow direction; the historical disaster data includes disaster type, occurrence time, occurrence scale, loss caused and frequency of disaster occurrence.

6. An intelligent geological disaster risk assessment system, used to implement an intelligent geological disaster risk assessment method as described in any one of claims 1 to 5, characterized in that: It includes data acquisition module, data preprocessing module, geological disaster risk assessment model building module, parameter search module and risk assessment module; The data acquisition module collects terrain data, geological data, meteorological data, hydrological data and historical disaster data, and sends the data to the data preprocessing module; The data preprocessing module receives the data sent by the data acquisition module, preprocesses the data, and sends the data to the geological disaster risk assessment model building module; The module for constructing a geological disaster risk assessment model receives data sent by the data preprocessing module, constructs a geological disaster risk assessment model module, and sends the data to the parameter search module and the risk assessment module; The parameter search module receives data sent by the module for building a geological disaster risk assessment model, searches for parameters of the model, and sends the data to the risk assessment module; The risk assessment module receives data sent by the geological disaster risk assessment model construction module and the parameter search module, collects terrain data, geological data, meteorological data and hydrological data of the assessment area, inputs the geological disaster risk assessment model, and the model outputs the geological disaster risk level of the assessment area.

Images