Landslide risk prediction method and system based on data intelligent analysis

Abstract

The invention belongs to the technical field of risk prediction, and discloses a landslide risk prediction method and system based on data intelligent analysis, and the method comprises the steps: obtaining corresponding multi-source heterogeneous information, carrying out the feature extraction and data fusion of the information, and obtaining a mountain monitoring basic data set; constructing a mountain twinborn feature model based on the mountain monitoring basic data set, and performing landslide scene simulation based on the mountain twinborn feature model to obtain corresponding risk feature information; constructing a trigger factor association network based on the risk feature information, and performing zoning evaluation and grading on the target landslide risk in combination with the mountain monitoring basic data set; constructing a corresponding grading risk zoning map; obtaining a landslide risk dynamic prediction result based on the grading risk zoning map; a corresponding grading early warning rule base is constructed; and the target mountain area is monitored in real time based on the grading early warning rule base, and the monitoring result is fed back. According to the invention, a comprehensive and efficient solution is provided for prevention and treatment of landslide disasters.

Description

Technical Field

[0001] The present invention relates to the technical field of risk prediction, and more specifically, to a landslide risk prediction method and system based on data intelligent analysis. Background Art

[0002] Landslide is a common geological disaster; with the continuous impact of human activities on the natural environment, the occurrence frequency and harm degree of landslides show an upward trend. Therefore, establishing an efficient and accurate landslide risk prediction system is of great significance for disaster prevention and mitigation.

[0003] Compared with the prior art, the following problems mainly exist in the traditional risk prediction process: Traditional landslide monitoring systems usually adopt single or limited data sources; the information obtained is limited and it is difficult to comprehensively reflect the actual state of the mountain body; existing landslide risk assessment methods are mostly based on empirical formulas or simplified models, and it is difficult to accurately simulate the landslide process under complex geological conditions; traditional warning systems usually use single indicators or simple threshold methods for warning, such as judging only based on cumulative rainfall or simple displacement rate; these warning indicators lack scientific basis, the threshold setting is mostly based on empirical judgment, and the same standard is often simply applied in different regions, resulting in low accuracy and reliability of the warning; In view of this, the present invention proposes a landslide risk prediction method and system based on data intelligent analysis to solve the above problems. Summary of the Invention

[0004] In order to overcome the above defects of the prior art and to achieve the above object, the present invention provides the following technical solutions: A landslide risk prediction system based on data intelligent analysis, comprising: A data acquisition module, configured to perform real-time monitoring and data collection on geological and meteorological parameters of a target mountain area, obtain corresponding multi-source heterogeneous information, and perform feature extraction and data fusion on the collected multi-source heterogeneous information to obtain a corresponding basic mountain monitoring data set; A risk simulation module, which constructs a mountain twin feature model based on the basic mountain monitoring data set, and performs multi-landslide scenario simulations based on it to obtain corresponding simulated risk data, and obtains risk characteristic information corresponding to different landslide scenarios according to the simulated risk data; A risk assessment module, which constructs a trigger factor association network based on the risk characteristic information, and combines the basic mountain monitoring data set to perform zonal assessment and grade division on the target landslide risk; constructs a corresponding graded risk zoning map; The risk prediction module conducts spatio-temporal evolution analysis and prediction analysis based on the hierarchical risk zoning map to obtain the corresponding dynamic prediction results of landslide risk; and based on the dynamic prediction results of landslide risk, it establishes an early warning index system and sets thresholds to obtain the corresponding hierarchical early warning rule base; and based on the hierarchical early warning rule base, it conducts real-time monitoring on the target mountain area and feeds back the monitoring results.

[0005] Furthermore, the process of obtaining the basic mountain monitoring data set includes: Conduct data verification and outlier screening on the obtained multi-source heterogeneous information; and unify the multi-source heterogeneous data after outlier screening, and the data unification includes time alignment and spatial registration; Conduct multi-source data fusion on the multi-source heterogeneous data after data unification, and conduct structured organization on the data after multi-source fusion to obtain the corresponding basic mountain monitoring data set.

[0006] Furthermore, the basic mountain monitoring data set includes three dimensions: time, space, and attribute; the time dimension records the observation time series, the space dimension includes geographical coordinates and elevation information, and the attribute dimension includes various monitoring parameters and derived indicators.

[0007] Furthermore, the process of obtaining risk characteristic information includes: Conduct three-dimensional modeling based on the basic mountain monitoring data set to obtain the mountain twin characteristic model corresponding to the target mountain area; Collect and organize historical landslide event data; and obtain known landslide scenarios based on it; Conduct landslide scenario simulation based on the mountain twin characteristic model to obtain simulated risk data under different types of landslide scenarios; Obtain the simulated risk data corresponding to several corresponding landslide scenarios; construct a two-dimensional rectangular coordinate system with each monitoring parameter in the simulated risk data as the vertical axis and time as the horizontal axis, and map the obtained simulated risk data into the two-dimensional rectangular coordinate system to obtain the corresponding risk evolution; extract the risk characteristic information in the risk evolution map, and the risk characteristic information includes parameter change trend, mutation characteristic, and critical point characteristic; Based on the above process of extracting risk characteristic information, obtain the risk characteristic information corresponding to other landslide scenarios.

[0008] Furthermore, the process of constructing the trigger factor correlation network includes: Extract monitoring indicators from the obtained historical landslide event data and basic mountain monitoring data to obtain the corresponding mountain monitoring indicators; and obtain the indicator characteristic information corresponding to each mountain monitoring indicator; Associate with risk characteristic information, conduct a relevance assessment of mountain monitoring indicators and landslide scenarios, and obtain the support degree and confidence degree between each mountain monitoring indicator and the corresponding landslide scenario; and based on this, obtain the correlation degree between each mountain monitoring indicator and the corresponding landslide scenario; Based on the obtained correlation degree, conduct a characteristic importance analysis of the corresponding mountain monitoring indicators to obtain the contribution degree of the mountain monitoring indicators to the corresponding landslide scenario; based on the obtained contribution degree, divide the corresponding mountain monitoring indicators into direct triggering factors and indirect triggering factors, and establish a hierarchical structure corresponding to the direct triggering factors and indirect triggering factors; Randomly construct several factor combinations based on the direct triggering factors and indirect triggering factors; Conduct a synergy analysis of the constructed factor combinations to obtain the corresponding synergy effect; and obtain the joint probability and conditional probability corresponding to each factor combination; Obtain the temporal dependence relationship between the triggering factors and the landslide scenario based on historical landslide event data; Construct a triggering factor association network based on the obtained contribution degree, hierarchical structure, synergy effect, and temporal dependence relationship; the triggering factors include direct triggering factors and indirect triggering factors.

[0009] Furthermore, the process of obtaining the hierarchical risk zoning map includes: Based on the constructed triggering factor association network and the mountain twin feature model, conduct regional division of the target mountain area to obtain several evaluation units; Conduct a sensitivity assessment of the divided evaluation units to obtain the sensitive evaluation index corresponding to the evaluation units; construct a coupling model based on the triggering factors and the sensitivity index, the model input of the coupling model includes the triggering factors and the sensitivity index involved in each evaluation unit, and the model output result of the coupling model is the instability probability of each evaluation unit; Based on the instability probability, conduct a risk level division of each evaluation unit to obtain the corresponding hierarchical risk zoning map.

[0010] Furthermore, the process of obtaining the dynamic prediction result of landslide risk includes: Conduct time series decomposition and trend analysis on the real-time collected mountain monitoring basic data set to obtain the corresponding time feature components, and based on the time feature components, construct a temporal prediction basic model, and integrate the constructed time prediction basic model to obtain the corresponding time evolution prediction framework; Based on the hierarchical risk zoning map, construct a spatial correlation model and a spatial association network, and through the spatial association network, conduct spatial diffusion simulation and boundary analysis to obtain a spatial evolution prediction framework; Integrate the constructed time evolution prediction framework and the spatial evolution prediction framework to obtain the corresponding mountain spatio-temporal prediction model; Based on the obtained spatio-temporal prediction model of the mountain body, scenario prediction is carried out to obtain the scenario prediction results corresponding to the target mountain body area, and confidence calculation is performed on the obtained scenario prediction results to obtain the corresponding confidence interval; Integrate the confidence interval with the corresponding scenario prediction results to obtain the corresponding dynamic prediction results of landslide risk.

[0011] Furthermore, the construction process of the hierarchical early warning rule base includes: Update the obtained triggering factors based on the obtained dynamic prediction results of landslide risk, obtain the synergy effect between the updated triggering factors, and construct the corresponding early warning index system based on it; Conduct data distribution analysis and threshold value research on the obtained early warning index system to obtain the corresponding preliminary index threshold scheme; Evaluate the effectiveness of the obtained initial index threshold scheme based on the obtained historical landslide event data; and select whether to adjust the initial index threshold scheme based on the effectiveness evaluation result. If not adjusted, output the initial index threshold scheme as the optimal index threshold scheme; if adjusted, adjust the corresponding initial index threshold scheme based on the preset adjustment strategy; Furthermore, conduct hierarchical division on the optimal index threshold scheme to obtain the early warning classification framework, and formulate different-level early warning triggering conditions and combination logics according to the early warning classification framework to obtain the hierarchical early warning rule base.

[0012] Furthermore, the process of real-time monitoring of the target mountain body area based on the hierarchical early warning rule base and feedback of the monitoring results includes: Extract the characteristics of the real-time collected mountain body basic monitoring data set, obtain the index characteristic information corresponding to each triggering factor, and conduct condition evaluation on it with the constructed hierarchical early warning rule base; perform early warning level conversion based on the condition evaluation result to obtain the corresponding early warning status vector; and generate the corresponding early warning signal and determine the comprehensive risk level based on it to obtain the corresponding mountain body early warning result and feedback it to the management department of the target mountain body area.

[0013] Furthermore, a method for predicting mountain landslide risk based on data intelligent analysis includes: Step 1: Real-time monitor and collect data on the geological and meteorological parameters of the target mountain body area to obtain the corresponding multi-source heterogeneous information, and extract the characteristics and fuse the data of the collected multi-source heterogeneous information to obtain the corresponding basic mountain body monitoring data set; Step 2: Construct a mountain body twin feature model based on the basic mountain body monitoring data set, and conduct multi-landslide scenario simulation based on it to obtain the corresponding simulated risk data, and obtain the risk characteristic information corresponding to different landslide scenarios according to the simulated risk data; Step 3: Construct a trigger factor association network based on the risk characteristic information, and combine it with the basic mountain monitoring dataset to conduct zonal assessment and grade division of the target landslide risk; construct a corresponding graded risk zoning map; Step 4: Conduct spatio-temporal evolution analysis and prediction analysis based on the graded risk zoning map to obtain the corresponding dynamic landslide risk prediction results; and establish an early warning index system and set thresholds based on the dynamic landslide risk prediction results to obtain the corresponding graded early warning rule base; conduct real-time monitoring on the target mountain area based on the graded early warning rule base and provide feedback on the monitoring results.

[0014] The technical effects and advantages of a landslide risk prediction method and system based on data intelligent analysis of the present invention: Through constructing a mountain twin feature model to realize multi-landslide scenario simulation, and combining with the extraction of risk characteristic information to establish a trigger factor association network; then conduct refined zonal assessment on the target mountain area to obtain a graded risk zoning map; further construct a prediction model based on spatio-temporal evolution analysis to obtain dynamic landslide risk prediction results; finally establish a scientific early warning index system and a graded early warning rule base to realize real-time monitoring and early warning, improve the prediction accuracy and early warning timeliness, reduce disaster losses, optimize the allocation of emergency resources, realize the full-cycle intelligent management of landslide risks, and provide a systematic solution for geological disaster prevention and control. Brief Description of the Drawings

[0015] Figure 1 It is a schematic diagram of a landslide risk prediction system based on data intelligent analysis of the present invention; Figure 2 It is a schematic diagram of a landslide risk prediction method based on data intelligent analysis of the present invention. Detailed Embodiments

[0016] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0017] Embodiment 1

[0018] Please refer to Figure 1 As shown, a landslide risk prediction system based on data intelligent analysis in this embodiment includes: A data acquisition module, which is used to monitor and collect geological and meteorological parameters of the target mountain area in real time, obtain corresponding multi-source heterogeneous information, and perform feature extraction and data fusion on the collected multi-source heterogeneous information to obtain a corresponding basic mountain monitoring data set; A risk simulation module, which constructs a mountain twin feature model based on the basic mountain monitoring data set, performs multi-landslide scenario simulations based on it, obtains corresponding simulated risk data, and obtains risk characteristic information corresponding to different landslide scenarios according to the simulated risk data; A risk assessment module, which constructs a trigger factor association network based on the risk characteristic information, and combines the basic mountain monitoring data set to conduct a zonal assessment and grade division of the target mountain landslide risk; constructs a corresponding graded risk zoning map; A risk prediction module, which conducts spatio-temporal evolution analysis and prediction analysis based on the graded risk zoning map to obtain corresponding dynamic landslide risk prediction results; and based on the dynamic landslide risk prediction results, establishes an early warning index system and sets thresholds to obtain a corresponding graded early warning rule base; based on the graded early warning rule base, conducts real-time monitoring of the target mountain area and provides feedback on the monitoring results; Each module is connected by wired and / or wireless means to achieve data transmission between modules.

[0019] It should be further noted that in the specific implementation process, the process of obtaining multi-source heterogeneous information includes: Using a distributed sensor network to collect geological parameter data in the target mountain area to obtain a corresponding basic geological data set, where the basic geological data set includes geological survey data, topographic and geomorphic data, and soil property data; among them, the distributed sensor network is composed of multiple sensing terminals buried at different depths and positions, and the sensing terminals include soil moisture content sensing terminals, pore water pressure sensing terminals, and displacement monitoring sensing terminals, and the sensing terminals are networked wirelessly to achieve timed data collection and transmission; Obtaining meteorological monitoring data corresponding to the target mountain area through local meteorological stations, where the meteorological monitoring data includes meteorological parameters such as rainfall, temperature change, and humidity change; Based on collection methods such as satellite remote sensing images and UAV aerial photography data, obtaining remote sensing images corresponding to the target mountain area, and performing image recognition on them to obtain corresponding topographic data, where the topographic data includes static parameters such as slope, aspect, and vegetation coverage rate, and dynamic parameters such as surface deformation; Summarizing the collected remote sensing images, topographic data, basic geological data set, and meteorological monitoring data to obtain corresponding multi-source heterogeneous information; One embodiment of the present invention includes, for example, installing inclination sensors and groundwater level sensors in boreholes within a mountain body to monitor internal deformation and groundwater level fluctuations; deploying soil moisture sensors and rainfall sensors on the slope surface to monitor rainfall infiltration; installing stress and strain sensors near potential slip surfaces to monitor the stress state of rock and soil masses; deploying an automatic weather station to collect meteorological data such as rainfall, temperature, and humidity; and simultaneously integrating high-resolution remote sensing image data to obtain large-scale surface coverage and deformation information.

[0020] It should be further noted that in the specific implementation process, the process of obtaining the basic dataset for mountain body monitoring includes: Data verification and outlier screening are performed on the obtained multi-source heterogeneous information; among them, the data verification includes data format verification and data integrity verification; the process of outlier screening includes: using the 3σ rule and combining with the parameter threshold range set by expert experience to judge the data parameters in the multi-source heterogeneous information to obtain corresponding potential outliers; for example, when the soil moisture content increases by more than 20% within a short period of time and there is no rainfall record, it can be marked as a potential outlier; Furthermore, the deviation degree between the corresponding potential outliers and the corresponding parameter threshold range is obtained, and different processing measures are taken based on the deviation degree; One embodiment of the present invention includes: if the deviation degree is less than 5%, the linear interpolation algorithm can be used to quickly repair the potential outliers; if the deviation degree is not less than 5%, different outlier processing strategies are adopted based on the corresponding data types at the potential outliers; for example: for data with temporal correlation such as displacement and moisture content, the ARIMA time series prediction algorithm is used for repair; for data with strong spatial correlation, such as surface temperature distribution, the Kriging spatial interpolation method is used for data completion; Due to the deviation in data upload frequency and collection location during the data collection process of different data sources, therefore, the multi-source heterogeneous data after outlier screening is unified, and the data unity refers to unifying the data from different sources into the standard spatio-temporal reference range, which is determined in advance by those skilled in the art based on industry standards; among them, the data unity includes time alignment and spatial registration, and the time alignment solves the problem of data time asynchrony caused by the deviation in data upload frequency by using the nearest time matching algorithm; spatial registration refers to ensuring the spatial consistency of data at different spatial positions during analysis by using coordinate transformation and unified geographical reference; Perform multi-source data fusion on the multi-source heterogeneous data after data unification, and perform structured organization on the data after multi-source fusion to construct a basic mountain monitoring dataset including three dimensions of time, space, and attributes; among them, the time dimension records the observation time series, the space dimension includes geographical coordinates and elevation information, and the attribute dimension includes various monitoring parameters and derived indicators, and the derived indicators represent composite indicators composed of multiple monitoring parameters, such as: the product of slope and soil moisture content (characterizing slope saturation), the ratio of cumulative rainfall to permeability coefficient (characterizing infiltration capacity), the comprehensive index of displacement rate and acceleration (characterizing the instability trend), etc.; It should be further noted that the multi-source data fusion integrates the same type of data in the multi-source heterogeneous data through a weighted average algorithm to improve data reliability. For example: the displacement data measured by GNSS and InSAR technologies can be dynamically weighted and fused according to weather conditions and observation angles to obtain a more reliable displacement measurement result.

[0021] It should be further noted that in the specific implementation process, the process of obtaining simulated risk data includes: Perform three-dimensional modeling based on the basic mountain monitoring dataset to obtain a mountain twin feature model corresponding to the target mountain area; among them, three-dimensional modeling refers to the process of constructing a digital twin model (i.e., mountain twin feature model) corresponding to the target mountain area by applying 3D modeling software and digital twin technology. The corresponding three-dimensional modeling process is prior art and will not be elaborated in this invention; Collect and organize historical landslide event data, which includes the time, location, scale, type of landslide, and environmental conditions before the occurrence; the sources of historical landslide event data include local chronicles, historical disaster archives, scientific research literature, and field investigation reports, etc.; Obtain known landslide scenarios based on the historical landslide event data, and the landslide scenarios include rainfall-induced landslides, earthquake-induced landslides, progressive landslides, etc.; Furthermore, perform landslide scenario simulation based on the mountain twin feature model to obtain simulated risk data under different types of landslide scenarios; among them, the landslide scenario simulation refers to the process of simulating the occurrence process of different types of landslide scenarios by adjusting factors such as soil moisture content, geological structure, rainfall parameters, and earthquake parameters in the mountain twin feature model; for example: taking the rainfall-induced landslide as an example; the rainfall-induced landslide refers to the landslide phenomenon caused by the increase in mountain water content due to continuous rainfall, the increase in soil mass and the decrease in strength; in the mountain twin feature model, by adjusting the rainfall intensity and duration parameters, the change in the internal water content of the mountain under different rainfall conditions can be simulated, and the response changes of various monitoring parameters in the mountain area can be obtained, including water content change, pore water pressure change, displacement change, etc., and they are used as the simulated risk data of the rainfall-induced landslide; It should be further noted that in the specific implementation process, the process of obtaining risk characteristic information includes: Taking a certain landslide scenario as an example, obtain the simulated risk data corresponding to several corresponding landslide scenarios; Taking each monitoring parameter in the simulated risk data as the vertical axis and time as the horizontal axis, construct a two-dimensional rectangular coordinate system, and map the obtained simulated risk data into the two-dimensional rectangular coordinate system to obtain the corresponding risk evolution diagram, and the risk evolution diagram is a curve; Furthermore, extract the risk characteristic information in the risk evolution diagram. The risk characteristic information includes parameter change trend, mutation characteristic, and critical point characteristic. Among them, the parameter change trend refers to the overall upward or downward direction of each monitoring parameter in the risk evolution diagram, indicating the overall change trend of the monitoring parameter over time; the mutation characteristic refers to the sharp change points of the monitoring parameters in the risk evolution diagram, indicating that the monitoring parameters change significantly at specific time points; the critical point characteristic refers to the inflection point where the change mode of the monitoring parameter in the risk evolution diagram changes from one state to another, indicating the critical moment when the mountain state undergoes qualitative change; Furthermore, based on the above process of extracting risk characteristic information, obtain the risk characteristic information corresponding to different types of landslide scenarios.

[0022] It should be further noted that in the specific implementation process, the process of constructing the trigger factor association network includes: Based on the principal component analysis algorithm, extract monitoring indicators from the obtained historical landslide event data and basic mountain monitoring data to obtain the corresponding mountain monitoring indicators. The mountain monitoring indicators include topographic and geological characteristics such as slope, aspect, elevation, lithology, and fault distribution, and also include rainfall intensity, cumulative rainfall, soil water content change rate, surface displacement rate, etc.; and obtain the index characteristic information corresponding to each mountain monitoring indicator. The index characteristic information includes the parameter change trend, mutation characteristic, and critical point characteristic of each mountain monitoring indicator in the basic mountain monitoring data; Combine risk characteristic information, and use the association rule mining algorithm to conduct a correlation assessment between the corresponding mountain monitoring indicators and the corresponding landslide scenarios, and obtain the support and confidence between each mountain monitoring indicator and the corresponding landslide scenario; and based on this, obtain the degree of correlation between each mountain monitoring indicator and the corresponding landslide scenario. Among them, the higher the support and confidence, the higher the corresponding degree of correlation; the association rule mining algorithm adopted in the present invention is the FP-growth algorithm; the said correlation assessment refers to setting the minimum support and confidence thresholds based on the FP-growth algorithm, extracting strong correlation rules, and combining time-series correlation analysis and spatial correlation analysis to obtain the correlation patterns at different times and spatial correlation patterns, and quantifying them as support and confidence; among them, the time-series correlation analysis uses the sliding time window method to identify the correlation patterns at different time lags; the spatial correlation analysis uses the geographically weighted statistical method to identify the spatial correlation patterns; Furthermore, based on the obtained degree of correlation, conduct a feature importance analysis on the corresponding mountain monitoring indicators to obtain the contribution degree of the mountain monitoring indicators to the corresponding landslide scenarios; the said feature importance analysis adopts the permutation importance method and combines the degree of correlation to quantify the contribution of each mountain monitoring indicator to the prediction of the occurrence probability of the landslide scenario, and the value range of the contribution degree is 0-100; Furthermore, based on the obtained contribution degree, divide the corresponding mountain monitoring indicators into direct triggering factors and indirect triggering factors, and establish a hierarchical structure corresponding to the direct triggering factors and indirect triggering factors; the said direct triggering factors usually include factors such as rainfall intensity, cumulative rainfall, seismic acceleration, etc. that can directly trigger landslides; the indirect triggering factors include internal factors such as slope, lithology, vegetation coverage, etc. that affect landslide sensitivity; the said hierarchical result is represented by a directed acyclic graph, which can be used to reflect the dependence relationship and mutual influence relationship between the direct triggering factors and the indirect triggering factors; Furthermore, randomly construct several factor combinations based on the direct triggering factors and indirect triggering factors. The factor combinations can be composed of two direct triggering factors or indirect triggering factors, or can be composed of one direct triggering factor and one indirect triggering factor; Conduct a synergy analysis on the constructed factor combinations based on the second-order interaction test method to obtain the corresponding synergy effect. The said synergy effect includes positive effect and negative effect; for example: high rainfall and steep slopes in the mountain area significantly increase the landslide risk, then the synergy effect of the corresponding factor combination is a negative effect; and obtain the joint probability and conditional probability corresponding to each factor combination. The said joint probability refers to the probability of the corresponding factor combination occurring within the landslide scenario obtained based on the historical landslide event data; the said conditional probability refers to the change in the occurrence probability of the landslide scenario under the corresponding factor combination; for example, the increase in the occurrence probability of the landslide scenario under the factor combination of high rainfall and high steep slopes; Perform a time-dependence analysis on the triggering factors based on the historical landslide event data to obtain the corresponding time-series dependence relationship, which represents the temporal relationship between the sequence of changes in the corresponding triggering factors and the occurrence time of the landslide scenario; the time-dependence analysis refers to the process of identifying the time-series patterns of the triggering factors and the occurrence of the landslide scenario through the Granger causality test algorithm; among them, the triggering factors include direct triggering factors and indirect triggering factors. Furthermore, construct a triggering factor association network based on the obtained contribution degree, hierarchical structure, synergy effect, and time-series dependence relationship. The triggering factor association network is represented by a graph data structure. The nodes in the triggering factor association network represent various triggering factors, and the edges represent the association relationships between the triggering factors.

[0023] It should be further noted that in the specific implementation process, a zonal assessment and grade division of the target mountain landslide risk are carried out; the process of constructing the corresponding graded risk zoning map includes: Based on the constructed triggering factor association network and the mountain twin feature model, the target mountain area is divided into several evaluation units; among them, the area division is carried out by combining the geological unit method and the regular grid method; to ensure that the geological characteristics within each evaluation unit are relatively uniform; Perform a sensitivity assessment on the divided evaluation units to obtain the sensitive evaluation index corresponding to the evaluation units; the sensitivity assessment represents the process of quantitatively evaluating the geological sensitivity within the corresponding evaluation units by comprehensively considering the influence of inherent factors such as slope, lithology, and structural plane characteristics, and using a multi-factor weighted scoring method; the geological sensitivity evaluation index is used to quantitatively characterize the susceptibility of landslides within the mountain area, reflecting the potential sensitivity degree of the mountain area itself to landslide disasters without considering the triggering factors; among them, the higher the sensitivity index, the higher the corresponding sensitivity degree; Furthermore, construct a coupling model based on the triggering factors and the sensitivity index. The model input of the coupling model includes the triggering factors and the sensitivity index involved in each evaluation unit, and the model output result of the coupling model is the instability probability of each evaluation unit; the instability probability represents the probability of landslide instability occurring in the corresponding evaluation unit under the corresponding triggering factors and sensitivity index; Among them, the coupling model is constructed using machine learning methods such as logistic regression or support vector machine. The construction process is prior art and will not be elaborated in detail in the present invention; and the training data collected during the construction process of the coupling model are historical landslide event data and historical mountain monitoring basic data corresponding to the target mountain area in a safe state; Furthermore, based on the instability probability, the risk levels of each evaluation unit are divided to obtain the corresponding graded risk zoning map.

[0024] It should be further noted that in the specific implementation process, the process of constructing a spatio-temporal evolution model and a prediction analysis framework based on the hierarchical risk zoning map and obtaining the corresponding dynamic landslide risk prediction results includes: Perform time series decomposition and trend analysis on the real-time collected basic mountain monitoring data set to obtain the corresponding time feature components, and construct a basic time series prediction model based on the time feature components to obtain a time evolution prediction framework; the time series decomposition uses the STL decomposition method, which is used to decompose the time series data into a trend term, a seasonal term, and a residual term; Specifically, decompose the time series corresponding to the monitoring parameters in the real-time basic mountain monitoring data set into a trend term, a residual term, and a seasonal term based on the STL decomposition method; Furthermore, perform trend analysis on the obtained trend term, residual term, and seasonal term; and obtain the corresponding time feature components based on the trend analysis results. The time feature components include trend features (long-term change direction and rate), periodic features (seasonal fluctuations and cyclic patterns), mutation features (abnormal points and structural changes), and residual features (random fluctuation part) of the corresponding monitoring parameters, etc.; and each feature component includes the mean, variance, skewness, kurtosis, autocorrelation coefficient, etc. of the corresponding monitoring parameter; Furthermore, construct the corresponding basic time prediction model based on each feature component in the time feature components. For example: for data with an obvious linear trend, use the ARIMA model, and the model order automatically selects the optimal parameter combination through the AIC criterion; for data with obvious seasonality, use the SARIMA model to capture seasonal and holiday effects; for complex non-linear time series, use the LSTM deep learning model to automatically learn the long-term and short-term dependencies of the time series; among them, the construction process of the corresponding basic time prediction model is prior art, and this application will not elaborate too much; Integrate the constructed basic time prediction model to obtain the corresponding time evolution prediction framework; wherein, the time evolution prediction framework can be used to predict the monitoring parameters at different time scales; Based on the hierarchical risk zoning map and spatial interpolation technology, a spatial correlation model and a spatial association network are constructed, and spatial diffusion simulation and boundary analysis are carried out through the spatial association network to obtain a spatial evolution prediction framework; the spatial correlation model is constructed by using the Geographically Weighted Regression (GWR) method to describe the spatial dependence structure between various triggering factors; the spatial association network is presented in the form of a graph structure to represent the spatial association relationship between different evaluation units in the mountain area; the spatial diffusion simulation represents the process of simulating the propagation of landslide scenarios on the target mountain area based on the cellular automata method; and the diffusion process in the simulation process needs to consider constraints such as terrain, hydrology, and geology, such as the risk will not spread to higher places, but will accelerate along the water flow direction; the boundary analysis refers to identifying the spatial boundary and boundary barrier effect in the process of landslide risk diffusion; the spatial boundary includes natural boundaries (such as ridge lines, rivers, geological boundaries) and artificial boundaries (such as roads, dams, protection projects); the boundary barrier effect is evaluated according to the boundary type and intensity, and is divided into strong barrier (completely blocking the risk diffusion), medium barrier (significantly slowing down the diffusion speed), and weak barrier (slightly affecting the diffusion direction). It should be further noted that the spatial evolution prediction framework supports multiple spatial prediction functions, including: risk diffusion path prediction, dynamic change prediction of the influence range, prediction of the location of new landslides, and prediction of the expansion direction of existing landslides. Integrate the constructed time evolution prediction framework and spatial evolution prediction framework to obtain a mountain spatio-temporal prediction model with multi-level prediction functions. Furthermore, based on the obtained mountain spatio-temporal prediction model, scenario prediction is carried out to obtain the corresponding scenario prediction results for the target mountain area. The scenario prediction results include the occurrence probability of various landslide scenarios, as well as the corresponding scenario risk zoning map and potential influence range under the corresponding landslide scenarios; among them, the content reflected in the scenario risk zoning map is the probability of landslide risk occurring in each evaluation unit under the corresponding landslide scenario. Furthermore, based on the uncertainty quantification algorithm, the confidence level of the obtained scenario prediction results is calculated to obtain the corresponding confidence interval. Integrate the confidence interval with the corresponding scenario prediction results to obtain the corresponding dynamic landslide risk prediction results; the dynamic landslide risk prediction results include the time dimension, spatial dimension, and probability dimension. The time dimension is used to display the time series predicted at different time scales. The time scales include daily prediction, weekly prediction, and monthly prediction; the prediction results at the corresponding time points under each time scale include the risk level, occurrence probability, and confidence interval; the spatial dimension includes the hierarchical risk zoning map used to display the spatial distribution of risks; the probability dimension displays the uncertainty and possibility distribution of the prediction, including the probability values of each risk level, the width of the confidence interval, and the coefficient of variation.

[0025] It should be further noted that in the specific implementation process, the construction process of the hierarchical early warning rule library includes: Updating the obtained triggering factors based on the dynamically predicted results of landslide risks, and obtaining the synergistic effect among the updated triggering factors; Furthermore, constructing a corresponding early warning index system based on the triggering factors and the synergistic effect; Conducting data distribution analysis and critical value research on the obtained early warning index system to obtain a corresponding preliminary index threshold scheme; the data distribution analysis refers to obtaining the statistical distribution characteristics of each triggering factor before the landslide based on the historical monitoring basic data set; the historical monitoring basic data set consists of the historical mountain body monitoring basic data sets before the landslide scenario occurs and during the mountain body safety period; furthermore, comparing the statistical distribution characteristics corresponding to before the landslide scenario occurs and during the mountain body safety period, the risk threshold interval corresponding to the landslide risk can be obtained; the critical value research refers to finding the optimal early warning threshold point corresponding to each triggering factor by using the principle of maximizing the Youden index; the preliminary index threshold scheme includes the four-level thresholds corresponding to each triggering factor; for example: the four-level thresholds of the 24-hour cumulative rainfall are 50mm (blue), 80mm (yellow), 120mm (orange), and 180mm (red) respectively; the four-level thresholds of the displacement rate are 5mm / day (blue), 10mm / day (yellow), 20mm / day (orange), and 30mm / day (red) respectively; Furthermore, evaluating the effectiveness of the obtained initial index threshold scheme based on the obtained historical landslide event data, and the effectiveness evaluation comprehensively evaluates the practicability of the corresponding initial index threshold scheme by calculating the missed alarm rate, false alarm rate, and early warning lead time; If the effectiveness evaluation result meets the expected requirements, the corresponding one will be output as the optimal index threshold scheme; if the effectiveness evaluation result does not meet the expected requirements, the corresponding initial index threshold scheme will be adjusted based on the pre-set adjustment strategy; for example: reducing the index threshold with a high missed alarm rate and increasing the index threshold with a high false alarm rate; adjusting the threshold interval and other methods; Furthermore, hierarchically dividing the corresponding optimal index threshold scheme based on the multi-level early warning principle announced by the industry standard to obtain an early warning classification framework, and formulating different-level early warning trigger conditions and combination logics according to the early warning classification framework to obtain a hierarchical early warning rule library.

[0026] It should be further noted that in the specific implementation process, the process of real-time monitoring of the target mountain area based on the hierarchical early warning rule library and feedback of the monitoring results includes: Extract the features of the real-time collected mountain foundation monitoring data set, obtain the index feature information corresponding to each triggering factor, and perform conditional evaluation on it with the constructed hierarchical early warning rule base; the conditional evaluation includes single-condition evaluation and composite-condition evaluation; the single-condition evaluation refers to comparing the parameter value of the corresponding triggering factor with its corresponding four-level threshold, and the composite-condition evaluation determines the priority of each triggering factor according to the synergistic effect between the triggering factors and the index feature information; Furthermore, based on the conditional evaluation results, perform early warning level conversion to obtain the corresponding early warning state vector; the early warning state vector is used to describe the current early warning state of different evaluation units within the mountain area; and based on it, generate the corresponding early warning signal and determine the comprehensive risk level, obtain the corresponding mountain early warning result, and feedback it to the management department of the target mountain area, so as to take different emergency measures according to the mountain early warning result.

[0027] The present invention obtains multi-source heterogeneous data through a distributed sensor network, a weather station and remote sensing technology and performs fusion; then constructs a mountain digital twin model for multi-scenario landslide simulation; then establishes a triggering factor association network for regional risk assessment; finally realizes spatio-temporal evolution prediction and establishes a hierarchical early warning rule base. The beneficial effects of the system are reflected in: significantly improving data quality and reliability; realizing accurate risk simulation under complex geological conditions; establishing a scientific and reasonable regional early warning index system; having the ability of spatio-temporal dynamic prediction of landslide risk; reducing the false alarm and missed alarm rate and extending the early warning lead time, providing a scientific basis for disaster prevention and mitigation decision-making.

[0028] Embodiment 2

[0029] Please refer to Figure 2 As shown, for the parts not described in detail in this embodiment, refer to the description content of Embodiment 1. Provide a method for predicting landslide risk based on data intelligent analysis, including: Step 1: Real-time monitor and collect the geological and meteorological parameters of the target mountain area to obtain the corresponding multi-source heterogeneous information, and perform feature extraction and data fusion on the collected multi-source heterogeneous information to obtain the corresponding mountain monitoring basic data set; Step 2: Based on the mountain monitoring basic data set, construct a mountain twin feature model, and perform multi-landslide scenario simulation based on it to obtain the corresponding simulated risk data, and obtain the risk feature information corresponding to different landslide scenarios according to the simulated risk data; Step 3: Based on the risk feature information, construct a triggering factor association network, and combine the mountain monitoring basic data set to conduct regional assessment and grade division of the target mountain landslide risk; construct the corresponding hierarchical risk zoning map; Step 4: Conduct spatio-temporal evolution analysis and prediction analysis based on the hierarchical risk zoning map to obtain the corresponding dynamic prediction results of landslide risks; and establish an early warning index system and set thresholds based on the dynamic prediction results of landslide risks to obtain the corresponding hierarchical early warning rule base; conduct real-time monitoring on the target mountain area based on the hierarchical early warning rule base and provide feedback on the monitoring results.

[0030] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

[0031] It should be noted that in this article, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements but also includes other elements not explicitly listed, or further includes elements inherent to such process, method, article or device. Without further limitations, an element defined by the statement "including one..." does not exclude the existence of additional identical elements in the process, method, article or device including the said element.

[0032] In the description of the present invention, it should be understood that the terms "first", "second", etc. are only used for descriptive distinction and cannot be understood as indicating or implying relative importance.

[0033] In the description of the present invention, unless otherwise specified, the meaning of "a plurality" is two or more.

[0034] In the description of the present invention, the meaning of "several" is one or more, and the meaning of "a large number" is two or more.

[0035] In the description of this specification, the description referring to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0036] For the formulas in this specification, only the numerical values are calculated after dimensionless processing. The formulas are obtained by collecting a large amount of data and performing software simulations to obtain a formula that is closest to the actual situation. The preset parameters and threshold values in the formulas are set by those skilled in the art according to the actual situation.

[0037] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the claims and their equivalents.

Claims

1. A landslide risk prediction system based on data intelligent analysis, characterized in that, Including: A data acquisition module, which is used to monitor and collect geological and meteorological parameters of the target mountain area in real time, obtain corresponding multi-source heterogeneous information, extract features and fuse data of the collected multi-source heterogeneous information, and obtain a corresponding basic mountain monitoring data set; A risk simulation module, which constructs a mountain twin feature model based on the basic mountain monitoring data set, conducts multi-landslide scenario simulations based on it, obtains corresponding simulated risk data, and obtains risk characteristic information corresponding to different landslide scenarios according to the simulated risk data; A risk assessment module, which constructs a trigger factor association network based on the risk characteristic information, and combines the basic mountain monitoring data set to conduct a zonal assessment and grade division of the target mountain landslide risk; constructs a corresponding graded risk zoning map; A risk prediction module, which conducts spatio-temporal evolution analysis and prediction analysis based on the graded risk zoning map, and obtains corresponding dynamic landslide risk prediction results; And based on the dynamic landslide risk prediction results, establish an early warning index system and set thresholds to obtain a corresponding graded early warning rule base; Based on the graded early warning rule base, conduct real-time monitoring of the target mountain area and provide feedback on the monitoring results.

2. The landslide risk prediction system based on data intelligent analysis according to claim 1, wherein The process of obtaining the basic mountain monitoring data set includes: Conduct data verification and outlier screening on the obtained multi-source heterogeneous information; and unify the multi-source heterogeneous data after outlier screening, and the data unification includes time alignment and spatial registration; Conduct multi-source data fusion on the unified multi-source heterogeneous data, and conduct structured organization on the data after multi-source fusion to obtain a corresponding basic mountain monitoring data set.

3. The landslide risk prediction system based on data intelligent analysis according to claim 2, wherein, The basic mountain monitoring data set includes three dimensions: time, space, and attribute; the time dimension records the observation time series, the space dimension includes geographical coordinates and elevation information, and the attribute dimension includes various monitoring parameters and derived indicators.

4. The landslide risk prediction system based on data intelligent analysis according to claim 3, wherein, The process of obtaining the risk characteristic information includes: Conduct three-dimensional modeling based on the basic mountain monitoring data set to obtain a mountain twin feature model corresponding to the target mountain area; Collect and organize historical landslide event data; and obtain known landslide scenarios based on it; Conduct landslide scenario simulations based on the mountain twin feature model to obtain simulated risk data under different types of landslide scenarios; Obtain simulated risk data corresponding to several corresponding landslide scenarios; construct a two-dimensional rectangular coordinate system with each monitoring parameter in the simulated risk data as the vertical axis and time as the horizontal axis, and map the obtained simulated risk data into the two-dimensional rectangular coordinate system to obtain a corresponding risk evolution; extract the risk characteristic information in the risk evolution map, and the risk characteristic information includes parameter change trends, mutation characteristics, and critical point characteristics; Based on the above process of extracting risk characteristic information, obtain risk characteristic information corresponding to other landslide scenarios.

5. The landslide risk prediction system based on data intelligent analysis according to claim 4, characterized in that, The process of constructing the trigger factor association network includes: Extract monitoring indicators from the obtained historical landslide event data and basic mountain monitoring data to obtain corresponding mountain monitoring indicators; and obtain the index characteristic information corresponding to each mountain monitoring indicator; Combined with risk characteristic information, conduct a relevance assessment of the mountain monitoring indicators and landslide scenarios, and obtain the support degree and confidence degree between each mountain monitoring indicator and the corresponding landslide scenario; and based on this, obtain the correlation degree between each mountain monitoring indicator and the corresponding landslide scenario; Based on the obtained correlation degree, conduct a characteristic importance analysis of the corresponding mountain monitoring indicators, and obtain the contribution degree of the mountain monitoring indicators to the corresponding landslide scenario; based on the obtained contribution degree, divide the corresponding mountain monitoring indicators into direct triggering factors and indirect triggering factors, and establish a hierarchical structure corresponding to the direct triggering factors and indirect triggering factors; Randomly construct several factor combinations based on the direct triggering factors and indirect triggering factors; Conduct a synergy analysis of the constructed factor combinations to obtain the corresponding synergy effect; and obtain the joint probability and conditional probability corresponding to each factor combination; Based on the historical landslide event data, obtain the time series dependence relationship between the triggering factors and the landslide scenarios; Based on the obtained contribution degree, hierarchical structure, synergy effect and time series dependence relationship, construct a triggering factor association network; the triggering factors include direct triggering factors and indirect triggering factors.

6. The landslide risk prediction system based on data intelligent analysis according to claim 5, characterized in that, The process of obtaining the graded risk zoning map includes: Based on the constructed triggering factor association network and the mountain twin feature model, conduct regional division of the target mountain area to obtain several evaluation units; Conduct a sensitivity assessment of the divided evaluation units to obtain the sensitive evaluation index corresponding to the evaluation units; based on the triggering factors and the sensitivity index, construct a coupling model, the model input of the coupling model includes the triggering factors and the sensitivity index involved in each evaluation unit, and the model output result of the coupling model is the instability probability of each evaluation unit; Based on the instability probability, conduct a risk level division of each evaluation unit to obtain the corresponding graded risk zoning map.

7. The landslide risk prediction system based on data intelligent analysis according to claim 6, wherein, The process of obtaining the dynamic prediction result of landslide risk includes: Conduct time series decomposition and trend analysis on the real-time collected mountain monitoring basic data set to obtain the corresponding time feature components, and based on the time feature components, construct a time series prediction basic model, and integrate the constructed time prediction basic model to obtain the corresponding time evolution prediction framework; Based on the graded risk zoning map, construct a spatial correlation model and a spatial association network, and through the spatial association network, conduct spatial diffusion simulation and boundary analysis to obtain a spatial evolution prediction framework; Integrate the constructed time evolution prediction framework and spatial evolution prediction framework to obtain the corresponding mountain spatio-temporal prediction model; Based on the obtained mountain spatio-temporal prediction model, conduct scenario prediction to obtain the scenario prediction result corresponding to the target mountain area, and calculate the confidence degree of the obtained scenario prediction result to obtain the corresponding confidence interval; Integrate the confidence interval and the corresponding scenario prediction result to obtain the corresponding dynamic prediction result of landslide risk.

8. The landslide risk prediction system based on data intelligent analysis according to claim 7, characterized in that, The process of constructing the graded early warning rule base includes: Based on the obtained dynamic prediction result of landslide risk, update the obtained triggering factors, and obtain the synergy effect between the updated triggering factors, and based on this, construct the corresponding early warning index system; Conduct data distribution analysis and critical value research on the obtained early warning index system to obtain the corresponding preliminary index threshold scheme; Evaluate the effectiveness of the obtained initial index threshold scheme based on the obtained historical landslide event data; and select whether to adjust the initial index threshold scheme based on the effectiveness evaluation result. If not adjusted, output the initial index threshold scheme as the optimal index threshold scheme; if adjusted, adjust the corresponding initial index threshold scheme based on the pre-set adjustment strategy; Furthermore, conduct hierarchical division on the optimal index threshold scheme to obtain the early warning classification framework, and formulate early warning trigger conditions and combination logics at different levels according to the early warning classification framework to obtain the hierarchical early warning rule base.

9. A landslide risk prediction system based on data intelligent analysis according to claim 8, characterized in that, The process of real-time monitoring the target mountain area based on the hierarchical early warning rule base and feeding back the monitoring results includes: Extract the characteristics of the real-time collected mountain basic monitoring data set, obtain the index characteristic information corresponding to each triggering factor, and conduct condition evaluation with the constructed hierarchical early warning rule base; conduct early warning level conversion based on the condition evaluation result to obtain the corresponding early warning status vector; and generate the corresponding early warning signal and determine the comprehensive risk level based on it to obtain the corresponding mountain early warning result and feed it back to the management department of the target mountain area.

10. A method for predicting landslide risks based on data intelligent analysis, which is based on the landslide risk prediction system based on data intelligent analysis according to any one of claims 1 to 9, and is characterized in that, Including: Step 1: Real-time monitor and collect data on the geological and meteorological parameters of the target mountain area to obtain the corresponding multi-source heterogeneous information, and extract the characteristics and fuse the data of the collected multi-source heterogeneous information to obtain the corresponding mountain monitoring basic data set; Step 2: Build a mountain twin feature model based on the mountain monitoring basic data set, and conduct multi-landslide scenario simulations based on it to obtain the corresponding simulated risk data, and obtain the risk characteristic information corresponding to different landslide scenarios according to the simulated risk data; Step 3: Build a trigger factor association network based on the risk characteristic information, and conduct zonal evaluation and grade division on the landslide risk of the target mountain in combination with the mountain monitoring basic data set; build the corresponding hierarchical risk zoning map; Step 4: Conduct spatio-temporal evolution analysis and prediction analysis based on the hierarchical risk zoning map to obtain the corresponding dynamic prediction results of landslide risk; And establish an early warning index system and set thresholds based on the dynamic prediction results of landslide risk to obtain the corresponding hierarchical early warning rule base; Conduct real-time monitoring on the target mountain area based on the hierarchical early warning rule base and feed back the monitoring results.

Images