High cutting slope stability evaluation method based on multi-source data fusion

Abstract

The invention discloses a high cut slope stability evaluation method based on multi-source data fusion, and particularly relates to the technical field of data analysis. By integrating topographic data, geological exploration data, remote sensing images, monitoring sensor data and geophysical detection data, an inversion algorithm is adopted to automatically extract abnormal structure features such as shear wave velocity sudden change and resistivity valley, the abnormal structure features and geological profiles are subjected to cross validation to generate an underground structure risk map layer, and a fusion model is combined with machine learning and a mechanical algorithm to obtain an underground structure risk map layer. According to the method, slope stability indexes including the safety coefficient and the sliding probability are comprehensively output, intelligent identification and grading of potential sliding dangerous areas are achieved, finally, an evaluation result is visually displayed in a three-dimensional GIS platform in a layer superposition mode, the identification precision and evaluation scientificity of slope hidden risks are remarkably improved, and the method is suitable for popularization and application. And a more reliable decision basis is provided for early warning and disposal of a high-risk slope.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and particularly to a method for evaluating the stability of high cut slopes based on multi-source data fusion. Background Art

[0002] The evaluation of the stability of high cut slopes refers to the process of analyzing and evaluating the stability of slopes (i.e., cut slopes) formed during the excavation of expressways. By comprehensively analyzing factors such as the geological structure, soil and rock properties, slope gradient, rainfall, and earthquake of the slopes, it is judged whether there is a risk of geological disasters such as landslides and collapses under natural conditions or external disturbances, so as to provide a scientific basis for slope design, construction, and later maintenance, and ensure the safe operation of the highway.

[0003] The existing technologies have the following deficiencies:

[0004] During the process of evaluating the stability of slopes based on multi-source data fusion, there may be a situation where underground abnormal structures are not identified in time, resulting in serious distortion of the evaluation results. Specifically, in some areas with developed karst, it is difficult for conventional remote sensing and surface monitoring means to obtain information on abnormal structures such as underground cavities, fissures, or weak interlayers. If geological profile or geophysical exploration data with sufficient depth is not introduced into the data fusion process, these hidden structures will be misjudged as homogeneous bodies, thus causing the model to seriously overestimate the overall stability of the slope. For example, during the construction of a certain expressway, the model evaluation results showed that the slope was "stable", but during actual construction, due to cutting through an unrecognized karst cave, the slope instantly lost stability, resulting in a major landslide accident. Summary of the Invention

[0005] The purpose of the present invention is to provide a method for evaluating the stability of high cut slopes based on multi-source data fusion to solve the deficiencies in the background art.

[0006] To achieve the above purpose, the present invention provides the following technical solution: A method for evaluating the stability of high cut slopes based on multi-source data fusion, including:

[0007] Obtain the topographic data, geological exploration data, remote sensing image data, monitoring sensor data, and geophysical exploration data of the high cut slope;

[0008] Based on the geophysical exploration data, use an inversion algorithm to automatically extract the characteristics of underground abnormal structures, and cross-verify with the geological profile data to generate an underground structure risk layer; the characteristics of the underground abnormal structures include shear wave velocity mutation characteristics and resistivity low valley anomaly characteristics;

[0009] Among them, after analyzing the resistivity low valley anomaly characteristics, generate a resistivity low valley anomaly value. The acquisition method is: construct a feature set of resistivity profile data. Suppose there are N sampling points or pixel points, and each point contains a resistivity value , local mean , local standard deviation and adjacent gradient ; The feature vector is expressed as: ; Select the number of clusters K, and use the K-means algorithm to cluster the feature vectors of N points : The initial centroids are randomly selected or initialized using K-means optimization until the centroids converge or reach the maximum number of rounds: From the final K clusters, select the cluster with the lowest mean, and the cluster with the lowest average resistivity is regarded as the resistivity trough anomaly class, labeled as , and for each sampling point i, calculate the resistivity trough anomaly value: ; In the formula, VC is the resistivity trough anomaly value, represents the maximum value of the Euclidean distance calculated between it and the low resistivity cluster center among all sampling points , represents the feature vector of the jth sample point;

[0010] Input the standardized terrain, geology, monitoring, and underground anomaly structure risk layers into the fusion model. The fusion model is constructed based on the coupling of machine learning algorithms and mechanical models, and outputs slope stability indicators, including safety factor and sliding probability;

[0011] According to the output results of the fusion model, conduct a hierarchical assessment of slope stability, combine the location of the underground anomaly structure, identify potential sliding hazard areas, and visually display the identification results in a 3D GIS platform in the form of layer superposition.

[0012] Preferably, an inversion algorithm is used to automatically extract the characteristics of the underground anomaly structure. Specifically: Deploy multiple seismic receivers on the target slope, use an active source to excite seismic waves, and collect surface wave signals; Denoise, band-pass filter, gain adjustment, and normalization processing are performed on the collected original seismic signals, and frequency-wavenumber analysis technology or phase velocity-frequency spectrum analysis technology is used to extract multiple surface wave dispersion curves; Input the extracted dispersion curves into the MASW inversion algorithm to reconstruct the shear wave velocity distribution of each depth point in the longitudinal section of the slope, forming a two-dimensional or three-dimensional shear wave velocity model; Calculate the wave velocity difference between adjacent points in the shear wave velocity model to obtain the wave velocity gradient field: Set the shear wave velocity mutation threshold, and mark the area that meets the conditions as the shear wave velocity mutation area, that is, the candidate area of the potential weak structural plane. Project the coordinates and depth of the mutation area into a vector or raster layer as the shear wave velocity mutation structure layer.

[0013] Preferably, an electrode array is arranged in the target area, the multi-pole configuration method is adopted to collect apparent resistivity data, and the finite element inversion algorithm is used to perform inversion calculation on the apparent resistivity data to generate a two-dimensional or three-dimensional resistivity distribution model of the underground; the resistivity model is standardized, the resistivity value is scaled to the interval [0,1], the low-resistance block is identified by image segmentation, the identified low-resistance anomaly is projected into a spatial layer, and its geometric boundary, depth range and central point coordinates are extracted to generate a resistivity trough anomaly layer.

[0014] Preferably, the machine learning model module is used to convert the shear wave velocity mutation feature and the resistivity trough anomaly feature into a comprehensive feature vector, use the comprehensive feature vector as the input of the machine learning model, and use the machine learning model to predict the sliding probability label of each spatial unit with each group of comprehensive feature vectors as the prediction target, and use the minimization of the sum of the prediction errors of the sliding probability labels of all each spatial unit as the training target to train the machine learning model until the sum of the prediction errors reaches convergence and stop the model training, and determine the sliding probability of each spatial unit according to the model output result, where the machine learning model is a polynomial regression model.

[0015] Preferably, after analyzing the shear wave velocity mutation feature, a shear wave velocity mutation value is generated, and the acquisition method is as follows:

[0016] Obtain the two-dimensional or three-dimensional shear wave velocity data volume Vs(x,z) of the slope, divide the model into several spatial units, and calculate the shear wave velocity difference on each vertical section : ; where: i represents the horizontal direction grid index; j represents the depth direction index; represents the shear wave velocity value of the i,j-th unit; represents the vertical shear wave velocity mutation value at this depth position; calculate the shear wave velocity gradient on each horizontal section : ; Define the shear wave velocity mutation value as the weighted superposition of the mutation values in two directions.

[0017] Preferably, after performing dimensionless normalization processing on the sliding probability of each obtained spatial unit and the mechanical safety factor of each cell, perform weighted average summation calculation on them to obtain the comprehensive instability score of each unit.

[0018] Preferably, compare the comprehensive instability score of each obtained unit with the gradient standard threshold, the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and compare the comprehensive instability score of each unit with the first standard threshold and the second standard threshold respectively;

[0019] If the comprehensive instability score of each unit is greater than the second standard threshold, the corresponding slope is marked as a high-risk area, and a first-level warning signal is generated at this time;

[0020] If the comprehensive instability score of each unit is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, the corresponding slope is marked as a medium-risk area, and a second-level warning signal is generated at this time;

[0021] If the comprehensive instability score of each unit is less than the first standard threshold, the corresponding slope is marked as a low-risk area, and a third-level warning signal is generated at this time.

[0022] Preferably, the identified high-risk and medium-risk areas are subjected to spatial overlay analysis with the underground abnormal structure layer, and the units within the spatially overlapping or adjacent ranges are identified and marked as potential sliding hazard areas to provide decision-making references for managers.

[0023] In the above technical solution, the technical effects and advantages provided by the present invention are as follows:

[0024] 1. By introducing geophysical exploration data and an automatic inversion algorithm, the present invention breaks through the problem of difficult identification of underground abnormal structures in traditional slope stability assessment, can accurately extract the characteristics of shear wave velocity mutations and resistivity low valley anomalies, and generates a high-confidence underground structure risk layer in cooperation with geological section cross-validation. Further, the present invention standardizes the terrain, geology, monitoring, and underground structure data and inputs them into the fusion model. The model combines machine learning algorithms and mechanical analysis methods to achieve high-precision prediction of slope stability indicators, outputs quantitative indicators including sliding probability and safety factor, and improves the scientificity and robustness of the assessment.

[0025] 2. The present invention divides the slope into high, medium, and low risks through a hierarchical threshold judgment mechanism, and combines the spatial position of the underground abnormal structure to identify potential sliding hazard areas and trigger a hierarchical warning mechanism. The identification results are visually displayed in the form of layers superimposed on the 3D GIS platform, making the risk information expression more intuitive, providing an efficient and operable auxiliary decision-making tool for engineering managers, and significantly improving the initiative and reliability of high cut slope disaster warning. Description of the Drawings

[0026] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings.

[0027] Figure 1 It is the schematic diagram of the present invention.

[0028] Figure 2 This is a schematic diagram of a machine learning model. Specific implementation manners

[0029] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are some but not all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0030] For the embodiments, please refer to Figure 1 As shown, a method for evaluating the stability of high cut slopes based on multi-source data fusion in this embodiment includes:

[0031] Obtain the topographic data, geological exploration data, remote sensing image data, monitoring sensor data and geophysical exploration data of the high cut slope;

[0032] Based on the geophysical exploration data, use the inversion algorithm to automatically extract the characteristics of underground abnormal structures, and perform cross-verification with the geological section data to generate an underground structure risk layer;

[0033] Input the standardized topographic, geological, monitoring and underground abnormal structure risk layers into the fusion model. The fusion model is constructed by coupling a machine learning algorithm and a mechanical model, and outputs slope stability indicators, including the safety factor and the sliding probability;

[0034] According to the output result of the fusion model, conduct a hierarchical evaluation of the slope stability, combine the position of the underground abnormal structure, identify the potential sliding hazard area, and visually display the identification result in a 3D GIS platform in the form of layer superposition.

[0035] The topographic data is mainly used to describe the elevation, slope, aspect and spatial form of the slope, providing basic geometric structure information for stability analysis. Common acquisition methods include:

[0036] Unmanned aerial vehicle (UAV) aerial survey: By carrying a high-precision lidar (LiDAR) or multispectral camera, quickly obtain the 3D point cloud and orthophoto of the slope, and generate a high-resolution digital elevation model (DEM) or digital surface model (DSM).

[0037] GNSS measurement and total station measurement: Set up ground measurement points at key control points to obtain high-precision topographic coordinate information for correcting remote sensing data.

[0038] Topographic map vectorization: Combine topographic maps with a scale of 1:1000 or 1:2000, and perform digital extraction through a GIS system.

[0039] Geological exploration data is used to reveal the rock and soil layer structure, physical and mechanical properties, and hydrogeological conditions of the slope. The acquisition methods include:

[0040] Drilling and sampling: Boreholes are arranged to obtain rock and soil samples at different depths, and laboratory tests (such as triaxial shear tests and direct shear tests) are conducted to analyze parameters such as the shear strength, porosity, and water content of the soil.

[0041] In-situ tests: Such as the standard penetration test (SPT), cone penetration test (CPT), and vane shear test (VST), which are used to obtain the bearing capacity and shear strength indexes of the strata.

[0042] Geological logging: Describe the drill core, identify the characteristics of structural planes such as faults, fractures, and joints, and draw profiles in combination with the regional geological map.

[0043] Remote sensing image data is used to analyze the macroscopic deformation of the slope, changes in vegetation cover, drainage paths, etc. Common acquisition methods include:

[0044] Satellite remote sensing images: Multispectral or high-resolution satellites (such as Sentinel-2, WorldView, and high-resolution series) are used to achieve periodic monitoring.

[0045] InSAR technology (interferometric synthetic aperture radar): Use radar images to generate deformation interferograms and monitor the small displacement changes on the ground surface, which is suitable for the early identification of landslides.

[0046] Infrared remote sensing and thermal imaging: Identify the changes in aquifers or the locations of potential slip surfaces in areas with large surface temperature differences.

[0047] Monitoring sensor data can achieve real-time perception and dynamic feedback of the slope stability. The common types of sensors include:

[0048] Inclinometer: Installed on the surface or in boreholes of the slope to monitor the sliding direction and angle changes.

[0049] Displacement meters and inclinometers: Used to measure the displacement trend of the slip zone layer and support the monitoring of shallow and deep slides.

[0050] Piezometers and groundwater level gauges: Monitor the groundwater pressure and water level changes to reflect the changes in potential sliding forces.

[0051] Rain gauges and weather stations: Collect meteorological elements such as rainfall intensity, rainfall duration, temperature, and humidity to evaluate the inducing factors.

[0052] Wireless Internet of Things acquisition system: Sensor data is transmitted to the slope safety management platform through a wireless network, supporting remote access and automatic early warning.

[0053] Geophysical exploration data is used to identify subsurface hidden abnormal structures (such as karst caves, fault zones, weak interlayers), which is the key data source for improving the accuracy of assessment depth, including:

[0054] Seismic wave method (reflection wave / refraction wave): By artificially exciting the seismic source, recording the propagation behavior of waves in the subsurface medium, and inversely inferring the subsurface layer interfaces and abnormal bodies.

[0055] Geoelectric method (resistivity imaging): By measuring the resistivity changes in different subsurface regions with electrodes, identifying weak structures such as aquifers, cavities, and faults.

[0056] Gravity and magnetic methods: Used to identify regions with significant lithological differences, and assist in judging subsurface dissolution or fracture characteristics.

[0057] Ground Penetrating Radar (GPR): Suitable for shallow structure scanning, obtaining high-resolution images for detail supplementation.

[0058] Based on geophysical exploration data, an inversion algorithm is used to automatically extract the characteristics of subsurface abnormal structures and cross-validate them with geological section data to generate a subsurface structure risk layer, specifically including:

[0059] Obtain the original seismic waves, geoelectric resistivity, or radar data in the slope area. Perform preprocessing operations on the data such as filtering, denoising, and normalization. Calibrate the data acquisition points and spatial coordinates for spatial registration with geological section data.

[0060] Steps for identifying shear wave velocity mutation regions, specifically including:

[0061] Deploy multiple seismic receivers on the target slope, use an active source to excite seismic waves, and collect surface wave signals.

[0062] Perform denoising, band filtering, gain adjustment, and normalization on the collected original seismic signals to remove background noise and enhance signal quality.

[0063] Use frequency-wavenumber (f-k) analysis or phase velocity-frequency spectrum analysis techniques to extract multiple surface wave dispersion curves.

[0064] Input the extracted dispersion curves into the MASW inversion algorithm to reconstruct the shear wave velocity (Vs) distribution at each depth point in the longitudinal section of the slope, forming a two-dimensional or three-dimensional Vs model.

[0065] Calculate the wave velocity difference between adjacent points (vertically or horizontally) in the Vs model to obtain the wave velocity gradient field:

[0066] Set the shear wave velocity mutation threshold (for example: ΔVs > 200 m / s), and mark the regions that meet this condition as shear wave velocity mutation regions, that is, candidate regions for potential weak structural planes.

[0067] Project the coordinates and depths of the mutation zones into vector or raster layers as the shear wave velocity mutation structure layers for subsequent fusion modeling.

[0068] Steps for extracting resistivity low valley anomalies specifically include:

[0069] Deploy an electrode array in the target area and use the multi-pole configuration method (such as Wenner or Schlumberger) to collect apparent resistivity data.

[0070] Use the finite element inversion algorithm to perform inversion calculations on the apparent resistivity data to generate a two-dimensional or three-dimensional underground resistivity distribution model, and the model grid accuracy can be set to 1 - 2 meters.

[0071] Perform normalization processing on the resistivity model to scale the resistivity values to the [0, 1] interval to enhance the image segmentation effect.

[0072] Identify low-resistance blocks through image segmentation: Option 1: Otsu algorithm (automatic threshold method), convert the resistivity model into a grayscale image; use the Otsu method to calculate the optimal threshold T, and segment the image into low-resistance and high-resistance parts; extract the area where the resistivity value is less than T as the low-resistance anomaly area.

[0073] Option 2: K-means clustering: Use the resistivity values of each grid point and its neighborhood statistical features as inputs; set the number of clusters K = 3 - 5 and cluster the resistivity values; select the cluster category with the lowest average resistivity as the low-resistance anomaly body area.

[0074] Project the identified low-resistance anomalies into a spatial layer, and extract their geometric boundaries, depth ranges, and central point coordinates to generate a resistivity low valley anomaly layer.

[0075] Project the underground anomaly structure features obtained from inversion, including shear wave velocity mutation features and resistivity low valley anomaly features, into a unified spatial coordinate system (such as the UTM coordinate system). Construct a three-dimensional point cloud or raster model to spatially annotate the anomaly structure.

[0076] Extract key markers such as structural planes, weak layers, and aquifers from the borehole logging. Establish profile intersection points corresponding to the inversion feature points, and compare their depths, thicknesses, and properties to perform consistency comparison. If the position deviation < 2m, mark it as an abnormal structure; if the difference is large, assign a low confidence level and enter the subsequent correction process.

[0077] Output the confirmed abnormal structural features and their spatial distributions in the form of raster layers (GeoTIFF or NetCDF). The layer attributes include: abnormal type (shear wave velocity mutation / low resistivity anomaly body); depth range (e.g., 8–15 m); risk level (high / medium / low, based on the structure scale and its impact on stability); confidence score (based on cross-validation consistency); the layer can be overlaid on a GIS platform or a 3D slope model for subsequent evaluation.

[0078] When new geophysical data or drilling data are input, trigger the automatic comparison mechanism; update the spatial boundaries and risk levels of abnormal structures in the layer to achieve data-driven dynamic correction.

[0079] Input the standardized terrain, geology, monitoring, and underground abnormal structure risk layers into the fusion model. The fusion model is constructed based on the coupling of machine learning algorithms and mechanical models, and outputs slope stability indicators, including the safety factor and the sliding probability. Specifically, it includes:

[0080] To ensure the comparability and fusion of multi-source heterogeneous data, first perform standardized preprocessing on the following data to unify the spatial resolution, coordinate system, and data format:

[0081] The input data categories include: terrain data (DEM / DSM): slope, aspect, curvature, slope height difference, etc. Geology data: lithology category coding, shear strength parameters (c, φ), void ratio, interlayer shear strength difference. Monitoring data: displacement rate, seepage pressure change, groundwater level change, rainfall intensity time series of each monitoring point.

[0082] The data standardization methods include: numerical data: use Z-score standardization or Min-Max normalization. Categorical data (such as lithology): use one-hot encoding. Time series data (such as displacement or rainfall): use sliding window statistical feature extraction (such as mean, volatility, slope).

[0083] Use methods such as the simplified Bishop method and the Janbu method to calculate the safety factor (FS) in the traditional sense based on the input geotechnical parameters, slope angle, groundwater level, etc. Introduce the weak structure area from the underground structure risk layer as the sliding surface control factor in the model, and assign reduced shear strength parameters to this area. Output the mechanical FS value of each cell.

[0084] The machine learning model module is used to: convert the shear wave velocity mutation feature and the resistivity trough anomaly feature into a comprehensive feature vector, use the comprehensive feature vector as the input of the machine learning model, take predicting the sliding probability label of each spatial unit with each group of comprehensive feature vectors as the prediction target by the machine learning model, take minimizing the sum of prediction errors of the sliding probability labels of all each spatial unit as the training target, train the machine learning model until the sum of prediction errors reaches convergence and then stop the model training, and determine the sliding probability of each spatial unit according to the model output result. Among them, the machine learning model is a polynomial regression model.

[0085] Among them, after analyzing the shear wave velocity mutation feature, a shear wave velocity mutation value is generated, and the acquisition method is as follows:

[0086] Obtain the two-dimensional or three-dimensional shear wave velocity data volume Vs(x,z) of the slope, which is derived from seismic surface wave inversion (such as MASW). Divide the model into several spatial units (such as dividing by a 1m×1m grid).

[0087] Calculate the shear wave velocity difference on each vertical section (along the depth direction) : ; where: i represents the horizontal grid index; j represents the depth direction index; represents the shear wave velocity value of the i,j-th unit (unit: m / s); represents the vertical shear wave velocity mutation value at this depth position.

[0088] Calculate the shear wave velocity gradient on each horizontal section (along the longitudinal direction of the slope) : ; represents the change amplitude of the wave velocity in the horizontal direction at the same depth.

[0089] Define the shear wave velocity mutation value as the weighted superposition of the mutation values in two directions, set the shear wave velocity mutation recognition threshold, and extract the area that meets the conditions as the mutation zone: if the shear wave velocity mutation value is greater than the shear wave velocity mutation recognition threshold, mark it as the shear wave velocity mutation area; where: the shear wave velocity mutation recognition threshold can be set according to the regional experience, such as 200–300 m / s; the continuously satisfied grids can be aggregated into a potential structural surface belt to generate a spatial vector layer.

[0090] After analyzing the resistivity trough anomaly feature, a resistivity trough anomaly value is generated, and the acquisition method is as follows: construct a feature set of resistivity profile data. Suppose there are N sampling points or pixel points, and each point contains: resistivity value , local mean , local standard deviation and adjacent gradient ; the feature vector is expressed as: .

[0091] Select the number of clusters K, which is usually determined by one of the following methods:

[0092] Empirical method: Let K = 3 - 5, corresponding to low, medium, and high resistivity regions respectively;

[0093] Elbow Method: Calculate the sum of squared errors (SSE) for different K values and select the point where "the SSE does not decrease significantly";

[0094] Silhouette Coefficient method: Measure the separation and compactness of the clusters and select the K value corresponding to the maximum silhouette value.

[0095] Use the K - means algorithm to cluster the feature vectors of N points as follows:

[0096] Randomly select the initial centroids or use K - means and optimized initialization;

[0097] Iteratively perform the following two steps until the centroids converge or reach the maximum number of iterations:

[0098] ; ; where: is the cluster category to which the k - th point belongs, is the centroid vector of the k - th cluster;

[0099] From the final K clusters, select the cluster with the lowest mean value. The cluster with the lowest average resistivity is regarded as the resistivity trough anomaly class and is marked as . For each sampling point i, calculate the resistivity trough anomaly value: ; In the formula, VC is the resistivity trough anomaly value, represents calculating the maximum Euclidean distance between it and the low - resistivity cluster center among all sampling points , represents the feature vector of the j - th sample point; The higher the resistivity trough anomaly value, the closer it is to the anomaly center.

[0100] After dimensionless normalization of the sliding probability of each obtained spatial unit and the mechanical safety factor of each cell, perform weighted average summation calculation on them to obtain the comprehensive instability score of each unit.

[0101] Compare the comprehensive instability score of each obtained unit with the gradient standard thresholds. The gradient standard thresholds include the first standard threshold and the second standard threshold, and the first standard threshold is less than the second standard threshold. Compare the comprehensive instability score of each unit with the first standard threshold and the second standard threshold respectively;

[0102] If the comprehensive instability score of each unit is greater than the second standard threshold, the corresponding slope is marked as a high-risk area, and a first-level warning signal is generated at this time;

[0103] If the comprehensive instability score of each unit is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, the corresponding slope is marked as a medium-risk area, and a second-level warning signal is generated at this time;

[0104] If the comprehensive instability score of each unit is less than the first standard threshold, the corresponding slope is marked as a low-risk area, and a third-level warning signal is generated at this time.

[0105] It should be noted here that the risk of the first-level warning signal is higher than that of the second-level warning signal, and the risk of the second-level warning signal is higher than that of the third-level warning signal. Relevant personnel can make corresponding treatments according to different levels of warning signals.

[0106] Perform a spatial overlay analysis on the identified high-risk and medium-risk areas and the underground abnormal structure layer, identify the units within the spatially overlapping or adjacent ranges, and mark them as potentially sliding dangerous areas. This marking is not only based on the scoring results but also takes into account the actual geological basis of structural weakening, enhancing the reliability of the identification.

[0107] Finally, import the DEM, abnormal structure layer, comprehensive instability score layer, and the marked potentially sliding areas into a 3D GIS platform for visual display. The platform can realize functions such as 3D rotation, cross-section browsing, click query, and hierarchical control, intuitively presenting the slope morphology, the possibility of sliding paths, the positions of abnormal structures, and their spatial relationships with risk areas, providing decision-making references for management personnel.

[0108] The above formulas are all dimensionless and take their numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulations to get a formula closest to the real situation. The preset parameters in the formulas are set by technicians in this field according to the actual situation.

[0109] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wire or wirelessly (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer, or a data storage device such as a server or data center that includes one or more collections of available media. The available media can be magnetic media (such as floppy disks, hard disks, magnetic tapes), optical media (such as DVDs), or semiconductor media. The semiconductor media can be a solid-state drive.

[0110] It should be understood that the term "and / or" in this document is merely a description of the association relationship between associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. Here, A and B can be singular or plural. In addition, the character " / " in this document generally represents an "or" relationship between the associated objects before and after, but it may also represent an "and / or" relationship, which can be specifically understood by referring to the context. Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in this document can be implemented by electronic hardware, or by a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0111] As described above, the above are only the specific embodiments of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed in the present application can easily think of changes or substitutions, which should all be covered by the protection scope of the present application.

Claims

1. A method for evaluating the stability of high cut slopes based on multi-source data fusion, characterized in that: Including: Obtain the topographic data, geological exploration data, remote sensing image data, monitoring sensor data, and geophysical exploration data of the high cut slope; Based on the geophysical exploration data, use the inversion algorithm to automatically extract the underground abnormal structure features, and perform cross-validation with the geological profile data to generate the underground structure risk layer; the underground abnormal structure features include the shear wave velocity mutation feature and the resistivity trough anomaly feature; Among them, after analyzing the resistivity trough anomaly characteristics, a resistivity trough anomaly value is generated. The acquisition method is as follows: construct a feature set of resistivity profile data. Suppose there are N sampling points or pixel points, and each point contains a resistivity value , local mean , local standard deviation and adjacent gradient ; the feature vector is expressed as: ; select the number of clusters K, and use the K-means algorithm to cluster the feature vectors of N points : the initial centroids are randomly selected or initialized using K-means optimization until the centroids converge or reach the maximum number of iterations: from the final K clusters, select the cluster with the lowest mean. The cluster with the lowest average resistivity is regarded as the resistivity trough anomaly class, marked as , for each sampling point i, calculate the resistivity trough anomaly value: ; in the formula, VC is the resistivity trough anomaly value, represents calculating the maximum value of the Euclidean distance between it and the low-resistivity cluster center among all sampling points , represents the feature vector of the jth sample point; Input the standardized topographic, geological, monitoring, and underground abnormal structure risk layers into the fusion model. The fusion model is constructed by coupling the machine learning algorithm and the mechanical model, and outputs the slope stability indicators, including the safety factor and the sliding probability; According to the output result of the fusion model, conduct a hierarchical evaluation of the slope stability. Combine the location of the underground abnormal structure, identify the potential sliding hazard area, and visually display the identification result in a 3D GIS platform in the form of layer superposition.

2. The high cutting slope stability evaluation method based on multi-source data fusion according to claim 1, wherein: The method for automatically extracting the underground abnormal structure features by using the inversion algorithm is as follows: Deploy multiple seismic receivers on the target slope, use the active source to excite seismic waves, and collect surface wave signals; Denoise, band-pass filter, gain adjust, and normalize the collected original seismic signals, and use the frequency-wavenumber analysis technique or the phase velocity-frequency spectrum analysis technique to extract multiple surface wave dispersion curves; Input the extracted dispersion curves into the MASW inversion algorithm to reconstruct the shear wave velocity distribution of each depth point in the longitudinal section of the slope to form a two-dimensional or three-dimensional shear wave velocity model; Calculate the wave velocity difference between adjacent points in the shear wave velocity model to obtain the wave velocity gradient field: Set the shear wave velocity mutation threshold, and mark the area that meets the conditions as the shear wave velocity mutation area, that is, the candidate area of the potential weak structural plane. Project the coordinates and depth of the mutation area into a vector or raster layer as the shear wave velocity mutation structure layer.

3. A method for evaluating the stability of high cut slopes based on multi-source data fusion according to claim 2, characterized in that: Deploy an electrode array in the target area, use the multi-pole configuration method to collect apparent resistivity data, and use the finite element inversion algorithm to perform inversion calculation on the apparent resistivity data to generate a two-dimensional or three-dimensional underground resistivity distribution model; Standardize the resistivity model, scale the resistivity value to the [0,1] interval, identify the low-resistance block through image segmentation, project the identified low-resistance anomaly body into a spatial layer, and extract its geometric boundary, depth range, and central point coordinates to generate the resistivity trough anomaly layer.

4. A method for evaluating the stability of high cut slopes based on multi-source data fusion according to claim 1, characterized in that: The machine learning model module is used to convert the shear wave velocity mutation feature and the resistivity trough anomaly feature into a comprehensive feature vector, use the comprehensive feature vector as the input of the machine learning model, take predicting the sliding probability label of each spatial unit with each group of comprehensive feature vectors as the prediction target, and take minimizing the sum of the prediction errors of the sliding probability labels of all each spatial unit as the training target to train the machine learning model until the sum of the prediction errors reaches convergence and stop the model training. Determine the sliding probability of each spatial unit according to the model output result. Among them, the machine learning model is a polynomial regression model.

5. The high cutting slope stability evaluation method based on multi-source data fusion according to claim 4, characterized in that: After analyzing the shear wave velocity mutation feature, generate the shear wave velocity mutation value. The acquisition method is as follows: Obtain the two-dimensional or three-dimensional shear wave velocity data volume Vs(x,z) of the slope, divide the model into several spatial units, and calculate the shear wave velocity difference on each vertical section : ; where: i represents the horizontal grid index; j represents the depth direction index; represents the shear wave velocity value of the i,j-th unit; represents the vertical shear wave velocity mutation value at this depth position; calculate the shear wave velocity gradient on each horizontal section : ; define the shear wave velocity mutation value as the weighted superposition of the mutation values in two directions.

6. The high cutting slope stability evaluation method based on multi-source data fusion according to claim 1, characterized in that: After performing dimensionless normalization on the sliding probability of each spatial unit and the mechanical safety factor of each cell, a weighted average summation calculation is carried out on them to obtain the comprehensive instability score of each unit.

7. A method for evaluating the stability of high cut slopes based on multi-source data fusion according to claim 6, characterized in that: The comprehensive instability score of each obtained unit is compared with the gradient standard thresholds. The gradient standard thresholds include a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold. The comprehensive instability score of each unit is respectively compared with the first standard threshold and the second standard threshold; If the comprehensive instability score of each unit is greater than the second standard threshold, the corresponding slope is marked as a high-risk area, and a first-level warning signal is generated at this time; If the comprehensive instability score of each unit is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, the corresponding slope is marked as a medium-risk area, and a second-level warning signal is generated at this time; If the comprehensive instability score of each unit is less than the first standard threshold, the corresponding slope is marked as a low-risk area, and a third-level warning signal is generated at this time.

8. The method for evaluating the stability of a high cutting slope based on multi-source data fusion according to claim 7, characterized in that: The identified high-risk and medium-risk areas are subjected to spatial overlay analysis with the underground abnormal structure layer to identify the units within the spatially overlapping or adjacent ranges, which are marked as potentially sliding dangerous areas to provide decision-making references for managers.

Images