Side slope potential slip plane identification method based on airborne LiDAR point cloud multi-scale fusion

Abstract

The invention discloses a slope potential slip plane identification method based on airborne LiDAR point cloud multi-scale fusion, and relates to the technical field of geological disaster monitoring, and the method comprises the steps: collecting original point cloud data, carrying out the preprocessing of the original point cloud data, obtaining a pure earth surface point cloud, constructing a multi-scale digital elevation model sequence, and carrying out the recognition of a slope potential slip plane from the multi-scale digital elevation model sequence. Extracting topographic features to obtain a multi-scale topographic feature map; performing unsaturated seepage simulation on the geometric boundary of the micro-scale digital elevation model in the multi-scale digital elevation model sequence in combination with rainfall working conditions set based on historical rainfall records to obtain a pore water pressure distribution cloud picture, and fusing a multi-scale topographic feature map to obtain a geometric slip sensitivity index; and coupling the pore water pressure distribution cloud picture and the geometric slip sensitivity index to generate a critical slip plane index picture. By visually outputting a slope stability analysis report, high-precision and quantifiable risk assessment of the potential slip plane of the slope is realized.

Description

Technical Field

[0001] This invention relates to the field of geological disaster monitoring technology, and in particular to a method for identifying potential slip surfaces on slopes based on multi-scale fusion of airborne LiDAR point clouds. Background Technology

[0002] With the increasing maturity of airborne lidar technology, its advantage of being able to quickly and accurately acquire large-scale three-dimensional point cloud data of the earth's surface has led to its widespread application in the field of slope geological hazard investigation. By analyzing terrain features at a single scale or a limited number of scales, combined with expert experience or machine learning algorithms, terrain features can be interpreted to delineate geometrically unstable areas. This method has greatly improved the efficiency and objectivity of landslide identification and provided a powerful technical tool for the preliminary screening of regional geological hazards.

[0003] The aforementioned methods, which rely on static geometric feature analysis, have an inherent limitation: they struggle to effectively couple the key external physical processes that trigger slope instability. In particular, the stability of slopes, particularly the core triggering mechanism of rainfall infiltration, is not solely determined by static topographic geometry but is also heavily influenced by dynamic hydrological conditions. Existing methods identify potential slip surfaces that primarily reflect slippery topographic features but fail to answer the crucial engineering question of under what rainfall conditions these slip surfaces will be activated. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a method for identifying potential slip surfaces on slopes based on multi-scale fusion of airborne LiDAR point clouds. This solves the problem that existing technologies cannot dynamically couple slope geometric features with key rainfall infiltration triggering mechanisms, thus failing to determine the activation risk of slip surfaces under specific rainfall conditions.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, the present invention provides a method for identifying potential slip surfaces of slopes based on multi-scale fusion of airborne LiDAR point clouds, which includes collecting raw point cloud data and preprocessing it to obtain a clean surface point cloud, constructing a multi-scale digital elevation model sequence, and extracting terrain features from the multi-scale digital elevation model sequence to obtain a multi-scale terrain feature map. The geometric boundaries of the mesoscale digital elevation model in the multi-scale digital elevation model sequence are combined with the rainfall conditions set based on historical rainfall records to conduct unsaturated seepage simulation, obtain pore water pressure distribution cloud map, fuse multi-scale topographic feature map to obtain geometric slip sensitivity index, and couple pore water pressure distribution cloud map and geometric slip sensitivity index to generate critical slip surface index map. Threshold segmentation and morphological processing are performed on the critical slip surface index map to identify and vectorize the potential slip surface boundary. Based on the point error distribution of the original point cloud data and the permeability coefficient range of the soil and rock mass in the unsaturated seepage numerical simulation, the Monte Carlo simulation method is used to generate potential slip surface boundary samples with different spatial morphologies, and then map them to the three-dimensional spatial coordinate system defined by the multi-scale digital elevation model sequence for statistical analysis, generating a three-dimensional probability distribution field with confidence intervals. Visualize the three-dimensional probability distribution field with confidence intervals and output a slope stability analysis report.

[0007] As a preferred embodiment of the slope potential slip surface identification method based on multi-scale fusion of airborne LiDAR point clouds in this invention, the method includes the following steps: collecting raw point cloud data, preprocessing it to obtain clean surface point clouds, and constructing a multi-scale digital elevation model sequence: Airborne LiDAR scanning was performed on the target slope to obtain raw point cloud data. The raw point cloud data was then denoised and filtered to obtain a clean surface point cloud. A multi-scale digital elevation model sequence is constructed based on clean surface point clouds through three-dimensional voxel downsampling. The multi-scale digital elevation model sequence includes digital elevation models at macro, meso, and meso scales.

[0008] As a preferred embodiment of the slope potential slip surface identification method based on multi-scale fusion of airborne LiDAR point clouds in this invention, the method includes the following steps: extracting terrain features from a multi-scale digital elevation model sequence to obtain a multi-scale terrain feature map. Slope and aspect features are extracted from macro-scale digital elevation models in a multi-scale digital elevation model sequence using a grid neighborhood elevation difference algorithm. From the mesoscale digital elevation model in the multi-scale digital elevation model sequence, the plan curvature and profile curvature features are extracted using quadratic surface fitting. From the mesoscale digital elevation model in the multi-scale digital elevation model sequence, the surface roughness and normal vector variability features are extracted using the local surface fitting method of principal component analysis. Multi-scale topographic feature maps are obtained by fusing slope and aspect characteristics, planar curvature and profile curvature characteristics, and surface roughness and normal vector variability characteristics.

[0009] As a preferred embodiment of the slope potential slip surface identification method based on multi-scale fusion of airborne LiDAR point clouds in this invention, the method includes: combining the geometric boundaries of the mesoscale digital elevation model in the multi-scale digital elevation model sequence with rainfall conditions set based on historical rainfall records to perform unsaturated seepage simulation, obtaining a pore water pressure distribution cloud map, and fusing the multi-scale topographic feature map to obtain a geometric slip sensitivity index, comprising the following steps: The triangular mesh surface of the spatial mesoscale digital elevation model in the multi-scale digital elevation model sequence is used as the geometric boundary input to the unsaturated seepage numerical solver. The unsaturated seepage numerical solver is loaded with rainfall conditions set based on historical rainfall records. The unsaturated seepage numerical solver is run to solve the Richards equation and obtain the pore water pressure distribution cloud map in the slope body. Normalize the slope features, aspect features, plan curvature features and profile curvature features, surface roughness features and normal vector variability features in multi-scale topographic feature maps; Adaptive weights are calculated based on the variances of normalized slope characteristics, aspect characteristics, plane curvature characteristics, profile curvature characteristics, surface roughness characteristics, and normal vector variability characteristics. The geometric slip sensitivity index is obtained by weighting and fusing the normalized slope features, aspect features, plane curvature features, profile curvature features, surface roughness features and normal vector variability features using adaptive weights.

[0010] As a preferred embodiment of the slope potential slip surface identification method based on airborne LiDAR point cloud multi-scale fusion of the present invention, the method includes the following steps: coupling pore water pressure distribution cloud map with geometric slip sensitivity index to generate critical slip surface index map: The shear strength parameters obtained by direct shear test of soil and rock mass of target slope are combined with the critical threshold of pore water pressure value in pore water pressure distribution cloud map. The pore water pressure value at each location in pore water pressure distribution cloud map is compared with the critical threshold. The dynamic weight coefficient of each location is calculated based on the comparison results. The critical slip surface index map is generated by multiplying the dynamic weighting coefficients by the geometric slip sensitivity index values ​​in the geometric slip sensitivity index map.

[0011] As a preferred embodiment of the slope potential slip surface identification method based on airborne LiDAR point cloud multi-scale fusion of the present invention, the method includes the following steps: thresholding and morphological processing of the critical slip surface index map to identify and vectorize the potential slip surface boundary: The Otsu's method is used to perform threshold segmentation on the critical slip surface index map, and regions with index values ​​higher than the segmentation threshold are extracted to obtain a binary image. Morphological closing operations are performed on binary images to connect adjacent discrete regions in the binary image and fill the holes inside the regions; Edge tracing is performed on the morphologically processed connected regions to obtain their contour coordinates. The contour coordinates are converted into vector polygons to obtain the potential slip surface boundary.

[0012] As a preferred embodiment of the slope potential slip surface identification method based on multi-scale fusion of airborne LiDAR point clouds in this invention, the method includes: based on the point error distribution of the original point cloud data and the permeability coefficient range of the soil and rock mass from unsaturated seepage numerical simulation, a Monte Carlo simulation method is used to generate potential slip surface boundary samples with different spatial morphologies, and these samples are mapped to a three-dimensional spatial coordinate system defined by a multi-scale digital elevation model sequence for statistical analysis, generating a three-dimensional probability distribution field with confidence intervals. This includes the following steps: A probability distribution model of point error is established based on the point error distribution of the original point cloud data, and a probability distribution model of permeability coefficient is established based on the permeability coefficient range of soil and rock mass in unsaturated seepage numerical simulation. The Monte Carlo simulation method was used to randomly select a set of parameter values ​​from the normal distribution model of the point error and the probability distribution model of the permeability coefficient, respectively. The original point cloud data is randomly perturbed using parameter values ​​to generate perturbed point cloud data. Based on the perturbed point cloud data, a multi-scale digital elevation model sequence is reconstructed using a three-dimensional voxelization downsampling method. Using parameter values ​​extracted from the probability distribution model of the permeability coefficient and combined with rainfall conditions set by historical rainfall records, a new set of pore water pressure distribution cloud maps is generated. Multi-scale topographic feature maps are extracted from the reconstructed multi-scale digital elevation model sequence, and the new pore water pressure distribution cloud map is fused with the multi-scale topographic feature maps to generate a new critical slip surface index map. Threshold segmentation and morphological processing are performed on the new critical slip surface index map to identify and vectorize a new potential slip surface boundary, and a set of potential slip surface boundary samples with different spatial morphologies is generated. Each potential slip surface boundary sample in the set of potential slip surface boundary samples with different spatial morphologies is uniformly mapped to the three-dimensional spatial coordinate system defined by the reconstructed multi-scale digital elevation model sequence. Spatial statistics are performed on the mapped potential slip surface boundary samples, and the kernel density estimation algorithm is used to calculate the frequency of sample occurrence at each spatial location. The frequency of sample occurrence is normalized to a probability value between zero and one, generating a three-dimensional probability distribution field with confidence intervals.

[0013] As a preferred embodiment of the slope potential slip surface identification method based on multi-scale fusion of airborne LiDAR point clouds in this invention, the following steps are included: visualizing the three-dimensional probability distribution field with confidence intervals and outputting a slope stability analysis report: Import the 3D probability distribution field with confidence intervals into the 3D rendering engine; In the 3D rendering engine, a probabilistic rendering threshold is set, and the spatial geometric parameters of high-risk areas are extracted based on the rendering results. The spatial geometric parameters of high-risk areas are then integrated with geological survey information to generate a slope stability analysis report.

[0014] In a second aspect, the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the slope potential slip surface identification method based on airborne LiDAR point cloud multi-scale fusion as described in the first aspect of the present invention.

[0015] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the slope potential slip surface identification method based on airborne LiDAR point cloud multi-scale fusion as described in the first aspect of the present invention.

[0016] The beneficial effects of this invention are as follows: By constructing a multi-scale digital elevation model sequence and extracting multi-scale topographic feature maps, the geometric boundary of the micro-scale digital elevation model is combined with rainfall conditions based on historical rainfall records to conduct unsaturated seepage simulation and obtain a pore water pressure distribution cloud map. Then, the multi-scale topographic features are integrated to obtain a geometric slip sensitivity index. By coupling the pore water pressure distribution cloud map and the geometric slip sensitivity index, a critical slip surface index map is generated. The Monte Carlo simulation method is used to quantify the positional error of the original point cloud data and the uncertainty of seepage simulation parameters, generating a set of potential slip surface boundary samples with different spatial morphologies. A three-dimensional probability distribution field with confidence intervals is generated through three-dimensional kernel density estimation and statistical analysis. A slope stability analysis report is output through visualization, realizing a high-precision and quantifiable risk assessment of potential slope slip surfaces. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of a method for identifying potential slip surfaces on slopes based on multi-scale fusion of airborne LiDAR point clouds.

[0019] Figure 2 A schematic diagram of the construction of a multi-scale digital elevation model sequence.

[0020] Figure 3 A flowchart for generating the geometric slip sensitivity index.

[0021] Figure 4 A flowchart for generating the critical slip surface index diagram is provided. Detailed Implementation

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

[0023] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0024] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0025] Reference Figures 1-4 This is one embodiment of the present invention, which provides a method for identifying potential slip surfaces of slopes based on multi-scale fusion of airborne LiDAR point clouds, including the following steps: S1. Collect raw point cloud data and preprocess it to obtain clean surface point cloud, and construct a multi-scale digital elevation model sequence.

[0026] S1.1. Airborne LiDAR scanning is performed on the target slope to obtain raw point cloud data. The raw point cloud data is then denoised and filtered to obtain a clean surface point cloud.

[0027] Furthermore, a flight path is planned for the target slope area, and an aerial platform equipped with lidar scans along the path. The laser pulses emitted by the lidar are reflected by the ground and received by the sensor, recording the three-dimensional coordinates and echo intensity of each laser point, thereby obtaining the original point cloud data covering the target slope. Subsequently, an outlier removal algorithm based on statistical analysis is applied to the original point cloud data to remove obvious noise points, and a slope-based filtering algorithm is used to distinguish between ground points and non-ground points, filtering out the non-ground points to obtain a clean surface point cloud that accurately represents the slope surface morphology.

[0028] S1.2. Based on the clean surface point cloud, a multi-scale digital elevation model sequence is constructed by three-dimensional voxelization downsampling. The multi-scale digital elevation model sequence includes digital elevation models at macro, meso, and meso scales.

[0029] Furthermore, based on obtaining a clean surface point cloud, a multi-scale digital elevation model sequence is constructed using a three-dimensional voxel downsampling method. The three-dimensional space containing the clean surface point cloud is divided into regular three-dimensional voxel grids. By setting different voxel sizes, for example, voxel grids with a side length of 5 meters are used to generate macro-scale digital elevation models, voxel grids with a side length of 1 meter are used to generate meso-scale digital elevation models, and voxel grids with a side length of 0.2 meters are used to generate meso-scale digital elevation models, thus forming a digital elevation model sequence that includes macro-scale, meso-scale, and meso-scale.

[0030] S2. Extract terrain features from the multi-scale digital elevation model sequence to obtain a multi-scale terrain feature map.

[0031] S2.1 Extract slope and aspect features from the macro-scale digital elevation model in the multi-scale digital elevation model sequence using the grid neighborhood elevation difference algorithm.

[0032] The slope characteristic expression is: ; in, Slope characteristics, for The difference in coordinates, In order to be in Horizontal distance in the direction, In order to be in Horizontal distance in the direction.

[0033] The characteristic expression for slope aspect is: ; in, This is a slope aspect characteristic.

[0034] Furthermore, starting from the macro-scale digital elevation model in the multi-scale digital elevation model sequence, a grid neighborhood elevation difference algorithm is used to extract slope and aspect features. Specifically, for each grid center point in the macro-scale digital elevation model, eight neighboring grid points are selected to form a 3x3 local window, and the elevation change rate of the center point in the east-west direction is calculated. The rate of change of elevation in the north-south direction, i.e. ;Will and Substituting the numerical values ​​into the slope characteristic expression Calculations are performed to obtain slope characteristic values, and then... and Substituting the numerical values ​​into the slope aspect characteristic expression Calculations are performed to obtain slope aspect characteristic values; all grid points in the macro-scale digital elevation model are traversed to complete the extraction of slope and aspect characteristics.

[0035] It should be noted that, at the physical level, Indicates elevation along The rate of change of direction (usually due east) is physically represented by the slope component in the east-west direction. It quantifies the degree and rate of change of the Earth's surface in the east-west direction. Indicates elevation along The rate of change of direction (usually referring to due north) is physically represented by the slope component in the north-south direction. It quantifies the degree and rate of change of the Earth's surface tilt in the north-south direction.

[0036] S2.2 Extract plane curvature and profile curvature features from the mesoscale digital elevation model in the multi-scale digital elevation model sequence using quadratic surface fitting.

[0037] Furthermore, starting from the mesoscale digital elevation model in the multi-scale digital elevation model sequence, a quadratic surface fitting method is used to extract planar curvature and profile curvature features. Specifically, for each grid point in the mesoscale digital elevation model, a local window is selected centered on that point, and a quadratic surface equation is fitted using the elevation values ​​of all grid points within the window. By calculating the coefficients of the quadratic surface equation, planar curvature feature values ​​reflecting the degree of contour line curvature and profile curvature feature values ​​reflecting the concavity and convexity changes along the slope direction are derived respectively. By traversing all grid points in the mesoscale digital elevation model, the extraction of planar curvature and profile curvature features is completed.

[0038] S2.3. From the microscale digital elevation model in the multiscale digital elevation model sequence, the surface roughness and normal vector variability features are extracted using the local surface fitting method of principal component analysis.

[0039] Furthermore, starting from the mesoscale digital elevation model within the multi-scale digital elevation model sequence, a local surface fitting method based on principal component analysis is used to extract surface roughness and normal vector variability features. For each point in the mesoscale digital elevation model, a local point set containing neighboring points is selected as the center. Principal component analysis is then performed on the local point set. Surface roughness features are obtained by calculating the standard deviation of the elevation values ​​of the local point set. Normal vector variability features are quantified by the ratio of the eigenvalues ​​of the first and third principal components obtained from the principal component analysis, reflecting the dispersion of the local surface normal vector direction. By traversing all points in the mesoscale digital elevation model, the extraction of surface roughness and normal vector variability features is completed.

[0040] S2.4. Multi-scale terrain features are obtained by fusing slope and aspect characteristics, planar curvature and profile curvature characteristics, and surface roughness and normal vector variability characteristics.

[0041] Furthermore, the slope and aspect features extracted from the macro-scale digital elevation model, the planar curvature and profile curvature features extracted from the meso-scale digital elevation model, and the surface roughness and normal vector variability features extracted from the micro-scale digital elevation model are fused together. The fusion process first normalizes each feature map to the same numerical range, and then superimposes the normalized slope, aspect, planar curvature, profile curvature, surface roughness, and normal vector variability features on each corresponding spatial grid point to generate a comprehensive multi-scale topographic feature map.

[0042] S3. Combine the geometric boundaries of the microscale digital elevation model in the multi-scale digital elevation model sequence with the rainfall conditions set based on historical rainfall records to conduct unsaturated seepage simulation, obtain the pore water pressure distribution cloud map, and integrate the multi-scale topographic feature map to obtain the geometric slip sensitivity index.

[0043] S3.1. The triangular mesh surface of the spatial microscale digital elevation model in the multiscale digital elevation model sequence is used as the geometric boundary input to the unsaturated seepage numerical solver.

[0044] Furthermore, the irregular triangular mesh surface of the microscale digital elevation model in the multi-scale digital elevation model sequence is extracted. The surface is composed of a large number of triangular patches, which accurately describes the micro-geomorphology of the slope surface. The irregular triangular mesh surface is used as the geometric boundary condition for the numerical simulation of unsaturated seepage and is input into the unsaturated seepage numerical solver to define the computational domain space for water transport in the soil and rock mass.

[0045] S3.2. Load the rainfall conditions based on historical rainfall records into the unsaturated seepage numerical solver, run the unsaturated seepage numerical solver to solve the Richards equation, and obtain the pore water pressure distribution cloud map in the slope body.

[0046] The Richards equation is: ; in, This refers to the volume of water per unit volume of rock or soil. For matrix suction, The unsaturated permeability coefficient, For Hami operators, For time, The symbol is for partial differentials. The coordinates are in the vertical direction.

[0047] Furthermore, in the unsaturated seepage numerical solver, characteristic rainfall event parameters obtained from the statistical analysis of historical rainfall records in the target slope area are loaded; initial values ​​of hydraulic parameters of the soil and rock mass are set; the unsaturated seepage numerical solver is run to solve the Richards equation describing water transport in the unsaturated zone, simulating the dynamic changes of pore water pressure in the slope body throughout the rainfall process; after the simulation is completed, the pore water pressure distribution cloud map in the slope body at a specific time or under specific conditions is output.

[0048] S3.3 Normalize the slope features, aspect features, planar curvature features, profile curvature features, surface roughness features, and normal vector variability features in the multi-scale topographic feature map.

[0049] Furthermore, the slope features, aspect features, plane curvature features, profile curvature features, surface roughness features, and normal vector variability features contained in the multi-scale topographic feature map are normalized respectively. The normalization process adopts the minimum-maximum scaling method, which linearly transforms the value of each feature at all grid points to the range of 0 to 1.

[0050] S3.4. Adaptive weights are calculated based on the variances of normalized slope characteristics, aspect characteristics, plane curvature characteristics, profile curvature characteristics, surface roughness characteristics, and normal vector variability characteristics.

[0051] The normalized variance expression is: ; in, The weight being calculated is currently being calculated. Variance of features The total number of grid points. For the first The feature in the first Normalized values ​​at each grid point For the first The average value of each feature at grid points For grid point indexing, For fixed feature indexes.

[0052] The adaptive weight expression is: ; in, For the first Adaptive weights for each feature, The total number of features, For variable feature index, In the summation of the first Variance of features.

[0053] Furthermore, based on the normalized values ​​of slope, aspect, plane curvature, profile curvature, surface roughness, and normal vector variability, the variance of each feature is calculated. The variance is calculated as the average of the sum of squared deviations of the normalized value of each feature from the mean across all grid points. An adaptive weight is then calculated based on the variance of each feature, which is the proportion of the variance of a single feature to the sum of the variances of all six features.

[0054] S3.5. Using adaptive weights, the normalized slope features, aspect features, plane curvature features, profile curvature features, surface roughness features, and normal vector variability features are weighted and fused to obtain the geometric slip sensitivity index.

[0055] Furthermore, adaptive weights are used to perform weighted summation on the normalized slope features, aspect features, plane curvature features, profile curvature features, surface roughness features, and normal vector variability features. For each grid point on the slope surface, the geometric slip sensitivity index value is equal to the sum of all six normalized feature values ​​multiplied by their respective adaptive weights.

[0056] S4. Couple the pore water pressure distribution cloud map with the geometric slip sensitivity index to generate the critical slip surface index map.

[0057] S4.1. The shear strength parameters obtained by the direct shear test of the target slope soil and rock mass are combined with the critical threshold of the pore water pressure value in the pore water pressure distribution cloud map. The pore water pressure value at each location in the pore water pressure distribution cloud map is compared with the critical threshold. The dynamic weight coefficient of each location is calculated based on the comparison results.

[0058] The expression for the critical threshold is: ; in, The critical threshold, The effective cohesion of the soil and rock mass, For safety reasons, The effective internal friction angle of the rock and soil mass. For the total normal stress, This represents the initial pore water pressure.

[0059] The expression for the dynamic weighting coefficient is: ; in, For dynamic weighting coefficients, This is the magnification factor. This represents the pore water pressure value at the current location. This represents the normalized range of pore water pressure variation.

[0060] Furthermore, by introducing a critical threshold based on the shear strength parameters of direct shear tests on soil and rock, and calculating dynamic weighting coefficients accordingly, the physical mechanism coupling between hydrological triggering conditions and the inherent anti-sliding capacity of the geological body is realized. The influence of pore water pressure is transformed from a static, uniform weighting distribution into a nonlinear, threshold-triggered dynamic response mechanism. When the pore water pressure does not exceed the critical threshold determined by the strength of the soil and rock, the slip surface identification is mainly dominated by geometric topographic features. Once the pore water pressure exceeds the critical threshold, the influence is significantly amplified through dynamic weighting, thereby accurately reflecting the area where the slope is most likely to become unstable under specific rainfall conditions. The method makes the identification results not only dependent on topography, but also more sensitive to key hydrological processes that lead to the deterioration of soil and rock strength, thus achieving a more physically consistent and predictive identification of potential slip surfaces.

[0061] It should be noted that, in multi-source information fusion, this step typically uses fixed empirical weights or statistically based weights (such as principal component analysis) to couple hydrological and geometric factors. This coupling method is essentially a mathematical superposition, lacking clear physical meaning and failing to reflect the nonlinear triggering effect of pore water pressure on slope stability. The fundamental difference between this step and existing technologies is that it abandons static, empirical weight allocation and instead uses a critical threshold with clear geotechnical meaning as a switch and regulator. The critical threshold is directly derived from the Mohr-Coulomb strength criterion, making the calculation of dynamic weight coefficients a simplified, physical-based stability assessment process. This method elevates information fusion from a mathematical combination to a physical mechanism-driven level, enabling the identification results to directly answer the engineering question of under what hydrological conditions the slip surface will be activated.

[0062] S4.2 Multiply the dynamic weighting coefficients by the geometric slip sensitivity index values ​​in the geometric slip sensitivity index diagram to generate the critical slip surface index diagram.

[0063] Furthermore, the geometric slip sensitivity index values ​​at the same spatial locations in the geometric slip sensitivity index map and the pore water pressure distribution cloud map are read; the geometric slip sensitivity index values ​​at the same spatial locations in the pore water pressure distribution cloud map are multiplied by the calculated dynamic weight coefficient, and the product is the index value of the same spatial location in the pore water pressure distribution cloud map in the critical slip surface index map; all spatial locations are traversed to complete the generation of the entire critical slip surface index map.

[0064] S5. Threshold segmentation and morphological processing are performed on the critical slip surface index map to identify and vectorize the potential slip surface boundary.

[0065] S5.1. The Otsu's method is used to perform threshold segmentation on the critical slip surface index map, and the regions in the critical slip surface index map whose index values ​​are higher than the segmentation threshold are extracted to obtain a binary image.

[0066] Furthermore, the critical slip surface index map is processed using the Otsu's method. The Otsu's method traverses all possible index values ​​in the critical slip surface index map as candidate segmentation thresholds. After dividing the image pixels into foreground and background classes using the candidate thresholds, the inter-class variance between the two classes is calculated. The candidate threshold that maximizes the inter-class variance is selected as the final segmentation threshold. Pixels in the critical slip surface index map with index values ​​higher than the segmentation threshold are marked as foreground, and pixels with index values ​​lower than the segmentation threshold are marked as background, thus obtaining a binary image.

[0067] S5.2 Perform morphological closing operations on the binary image to connect adjacent discrete regions in the binary image and fill the holes inside the regions.

[0068] Furthermore, morphological closing operations are performed on the binary image. The morphological closing operation is composed of a sequential combination of morphological dilation and morphological erosion operations. A structuring element is used to perform morphological dilation on the foreground region in the binary image, connecting adjacent discrete foreground regions and expanding the boundary of the foreground region. The same structuring element is then used to perform morphological erosion on the dilated image, restoring the approximate shape of the foreground region while filling the holes inside the region, resulting in a connected and complete binary foreground region.

[0069] S5.3 Perform edge tracking on the morphologically processed connected regions to obtain the contour coordinates of the connected regions.

[0070] Furthermore, edge tracking is performed on connected regions in the morphologically processed binary image. The edge tracking algorithm starts scanning from the top left corner of the binary image. When an unvisited foreground pixel is encountered, the algorithm uses this point as the starting point and tracks the outer contour of the connected region according to a certain neighborhood search rule. The coordinate sequence of all pixels on the contour is recorded. After completing the contour tracking of a connected region, the algorithm continues to scan the image to find the next unvisited foreground pixel. The above process is repeated until the entire binary image is traversed and the contour coordinate set of all connected regions is obtained.

[0071] S5.4. Convert the contour coordinates into vector polygons to obtain the potential slip surface boundary.

[0072] Furthermore, the set of contour coordinates obtained from edge tracking is converted into vector polygons; for the contour coordinate sequence of each connected region, adjacent points are connected by straight line segments in coordinate order to form a closed polygon; the closed polygon is the identified potential slip surface boundary, resulting in a set of potential slip surface boundaries represented by a series of vector polygons.

[0073] S6. Based on the point error distribution of the original point cloud data and the permeability coefficient range of the soil and rock mass in the numerical simulation of unsaturated seepage, the Monte Carlo simulation method is used to generate potential slip surface boundary samples with different spatial morphologies, and map them to the three-dimensional spatial coordinate system defined by the multi-scale digital elevation model sequence for statistical analysis, generating a three-dimensional probability distribution field with confidence intervals.

[0074] S6.1. Based on the point error distribution of the original point cloud data, establish a probability distribution model of the point error; based on the permeability coefficient range of the soil and rock mass in the unsaturated seepage numerical simulation, establish a probability distribution model of the permeability coefficient.

[0075] Furthermore, based on the calibration parameters of the airborne LiDAR sensor and the positioning and attitude determination accuracy of the flight operation, the three-dimensional coordinate error range of each laser foot point in the original point cloud data is determined. The error range is spatially independent and follows a normal distribution with a mean of zero. Then, the standard deviation of all point errors is used as the distribution parameter to establish a normal distribution model of the point error.

[0076] Based on the geological survey report of the target slope and the indoor geotechnical test data, the possible range of values ​​for the permeability coefficient of the rock and soil is determined. Since there is a lack of prior information in the logarithmic space and the permeability coefficient follows a uniform distribution, the logarithmic uniform distribution model of the permeability coefficient is established using the logarithmic value of the permeability coefficient range as the upper and lower bounds.

[0077] S6.2. Using the Monte Carlo simulation method, a set of parameter values ​​are randomly selected from the normal distribution model of the point error and the probability distribution model of the permeability coefficient, respectively.

[0078] Furthermore, the Monte Carlo simulation method is used to randomly select a set of point error parameter values ​​from the normal distribution model of point errors to disturb the original point cloud data; and to randomly select a set of permeability coefficient parameter values ​​from the probability distribution model of permeability coefficients to re-perform unsaturated seepage numerical simulation.

[0079] S6.3. Randomly perturb the original point cloud data using parameter values ​​to generate perturbed point cloud data. Based on the perturbed point cloud data, reconstruct a multi-scale digital elevation model sequence using a three-dimensional voxelization downsampling method. Using parameter values ​​extracted from the probability distribution model of the permeability coefficient and combined with the rainfall conditions set by historical rainfall records, generate a new pore water pressure distribution cloud map.

[0080] Furthermore, the original point cloud data is randomly perturbed using parameter values ​​extracted from the normal distribution model of point errors to generate perturbed point cloud data. Based on the perturbed point cloud data, a multi-scale digital elevation model sequence is reconstructed using a three-dimensional voxelization downsampling method. Using parameter values ​​extracted from the probability distribution model of permeability coefficients and combined with rainfall conditions set from historical rainfall records, unsaturated seepage numerical simulation is re-executed to generate a new set of pore water pressure distribution cloud maps.

[0081] S6.4 Extract multi-scale topographic feature maps from the reconstructed multi-scale digital elevation model sequence, and fuse the new pore water pressure distribution cloud map with the multi-scale topographic feature maps to generate a new critical slip surface index map.

[0082] Furthermore, multi-scale topographic feature maps are extracted from the reconstructed multi-scale digital elevation model sequence; the new pore water pressure distribution cloud map is fused with the multi-scale topographic feature map to generate a new critical slip surface index map.

[0083] S6.5. Threshold segmentation and morphological processing are performed on the new critical slip surface index map to identify and vectorize a new potential slip surface boundary, and a set of potential slip surface boundary samples with different spatial morphologies is generated.

[0084] Furthermore, the new critical slip surface index map is segmented using the Otsu's method to extract regions with index values ​​higher than the segmentation threshold and generate binary images. Morphological closing operations are performed on the binary images to connect adjacent discrete regions and fill internal holes. Edge tracking is performed on the processed connected regions to obtain contour coordinate sequences. The contour coordinate sequences are converted into vector polygons to obtain a new potential slip surface boundary.

[0085] A new set of parameter values ​​is randomly selected from the established point error normal distribution model and permeability coefficient probability distribution model. The newly selected point error parameters are used to perturb the original point cloud data to generate a new point cloud. A multi-scale digital elevation model sequence is reconstructed based on the point cloud. At the same time, the newly selected permeability coefficient parameters are used in conjunction with rainfall conditions to perform unsaturated seepage simulation to generate a new pore water pressure distribution cloud map. Then, topographic features are extracted from the reconstructed multi-scale digital elevation model sequence and fused with the new pore water pressure distribution cloud map to generate a new critical slip surface index map. The index map is subjected to threshold segmentation, morphological processing and vectorization to obtain a new potential slip surface boundary sample. The above process is repeated until the preset number of simulations (e.g., one thousand times) is reached to generate a set of potential slip surface boundary samples with different spatial morphologies for all samples.

[0086] S6.6 Map each potential slip surface boundary sample in the set of potential slip surface boundary samples with different spatial morphologies to the three-dimensional spatial coordinate system defined by the reconstructed multi-scale digital elevation model sequence. Perform spatial statistics on the mapped potential slip surface boundary samples and use the kernel density estimation algorithm to calculate the sample occurrence frequency at each spatial location.

[0087] The expression for kernel density estimation is: ; in, for The probability density function value at that location. For the total number of Monte Carlo simulations, For the first Potential slip surface boundary samples obtained from Monte Carlo simulations. For a three-dimensional point on the potential slip surface boundary sample, For Gaussian kernel function, For smoothing parameters, Let be the spatial coordinate vector of any point to be evaluated in the three-dimensional probability distribution field.

[0088] Furthermore, each potential slip surface boundary sample in the set of potential slip surface boundary samples with different spatial morphologies is uniformly mapped to the three-dimensional spatial coordinate system defined by the reconstructed multi-scale digital elevation model sequence; spatial statistics are performed on the mapped potential slip surface boundary samples, and the kernel density estimation algorithm is used to calculate the sample occurrence frequency at each spatial location; the kernel density estimation algorithm uses a Gaussian kernel function to calculate the frequency of occurrence at any point in three-dimensional space. probability density value equal to all Monte Carlo simulations In this case, all potential slip surface boundary samples All points on Through Gaussian kernel function Calculated points Sum of the contribution values ​​of distance, then divide by With smoothing parameters The product of cubes.

[0089] It should be noted that the Gaussian kernel function ,in , representing the standardized spatial distance, the Gaussian kernel function with points Centered on a point, a symmetrical bell-shaped distribution is formed in three-dimensional space, with weights varying with the point. and The decay of the distance between them decreases exponentially with increasing Euclidean distance, and the decay rate is determined by the smoothing parameter. control.

[0090] S6.7 Normalize the frequency of sample occurrence to a probability value between zero and one, and generate a three-dimensional probability distribution field with confidence intervals.

[0091] Furthermore, the frequency of each sample occurrence at each spatial location is normalized to a probability value between zero and one using a linear transformation method. This linear transformation divides the frequency of each sample occurrence at each spatial location by the maximum value among all spatial location frequency values, thus limiting the normalized probability value to between zero and one. The normalized probability value is then assigned to each corresponding grid cell in the three-dimensional spatial coordinate system, generating a continuous three-dimensional probability distribution field. The value of each grid cell in the three-dimensional probability distribution field represents the probability that the spatial location belongs to a potential slip surface. By extracting isosurfaces with specific probability values, a confidence interval for the potential slip surface at a specific confidence level can be defined, thereby generating a three-dimensional probability distribution field with confidence intervals.

[0092] S7. Visualize the three-dimensional probability distribution field with confidence intervals and output a slope stability analysis report.

[0093] S7.1 Import the 3D probability distribution field with confidence intervals into the 3D rendering engine.

[0094] Furthermore, the data file of the three-dimensional probability distribution field with confidence intervals is imported into the three-dimensional rendering engine in a standard three-dimensional mesh data format; each mesh cell in the three-dimensional probability distribution field stores the probability value of a spatial location belonging to a potential slip surface; a multi-scale digital elevation model sequence that matches the spatial coordinates of the three-dimensional probability distribution field is loaded into the three-dimensional rendering engine and overlaid as a terrain base for display.

[0095] S7.2 Set the probability rendering threshold in the 3D rendering engine, extract the spatial geometric parameters of high-risk areas based on the rendering results, integrate the spatial geometric parameters of high-risk areas with geological survey information, and generate a slope stability analysis report.

[0096] Furthermore, a probability rendering threshold is set in the 3D rendering engine. Based on the set threshold, the 3D rendering engine performs color mapping and transparency rendering on the 3D probability distribution field, making high-risk areas appear as bright, semi-transparent areas in the 3D terrain scene. Based on the rendering results, the spatial geometric parameters of the high-risk areas are extracted using a 3D region selection tool, including the volume, surface area, average depth, and spatial coordinates of the high-risk areas. The spatial geometric parameters of the high-risk areas are integrated and analyzed with the geological survey information of the target slope, such as the type of soil and rock mass, the orientation of structural surfaces, and groundwater conditions. A slope stability analysis report is generated, which includes the spatial distribution, geometric characteristics, risk level assessment, and stability evaluation and treatment recommendations based on geological conditions for the high-risk areas.

[0097] This embodiment also provides a computer device applicable to the slope potential slip surface identification method based on airborne LiDAR point cloud multi-scale fusion, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the slope potential slip surface identification method based on airborne LiDAR point cloud multi-scale fusion as proposed in the above embodiment.

[0098] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0099] This embodiment also provides a storage medium storing a computer program. When executed by a processor, the program implements the slope potential slip surface identification method based on multi-scale fusion of airborne LiDAR point clouds as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0100] In summary, this invention constructs a multi-scale digital elevation model sequence and extracts multi-scale topographic feature maps. It combines the geometric boundaries of the micro-scale digital elevation model with rainfall conditions based on historical rainfall records to conduct unsaturated seepage simulation and obtain pore water pressure distribution cloud maps. Then, it integrates multi-scale topographic features to obtain a geometric slip sensitivity index. By coupling the pore water pressure distribution cloud map and the geometric slip sensitivity index, it generates a critical slip surface index map. It uses the Monte Carlo simulation method to quantify the positional errors of the original point cloud data and the uncertainty of seepage simulation parameters, generating a set of potential slip surface boundary samples with different spatial morphologies. It generates a three-dimensional probability distribution field with confidence intervals through three-dimensional kernel density estimation and outputs a slope stability analysis report through visualization, thus achieving a high-precision and quantifiable risk assessment of potential slope slip surfaces.

[0101] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for identifying potential slip surfaces on slopes based on multi-scale fusion of airborne LiDAR point clouds, characterized in that: This includes collecting raw point cloud data, preprocessing it to obtain clean surface point clouds, constructing a multi-scale digital elevation model sequence, and extracting terrain features from the multi-scale digital elevation model sequence to obtain a multi-scale terrain feature map. The geometric boundaries of the mesoscale digital elevation model in the multi-scale digital elevation model sequence are combined with the rainfall conditions set based on historical rainfall records to conduct unsaturated seepage simulation, obtain pore water pressure distribution cloud map, fuse multi-scale topographic feature map to obtain geometric slip sensitivity index, and couple pore water pressure distribution cloud map and geometric slip sensitivity index to generate critical slip surface index map. Threshold segmentation and morphological processing are performed on the critical slip surface index map to identify and vectorize the potential slip surface boundary. Based on the point error distribution of the original point cloud data and the permeability coefficient range of the soil and rock mass in the unsaturated seepage numerical simulation, the Monte Carlo simulation method is used to generate potential slip surface boundary samples with different spatial morphologies, and then map them to the three-dimensional spatial coordinate system defined by the multi-scale digital elevation model sequence for statistical analysis, generating a three-dimensional probability distribution field with confidence intervals. Visualize the three-dimensional probability distribution field with confidence intervals and output a slope stability analysis report.

2. The method for identifying potential slip surfaces of slopes based on multi-scale fusion of airborne LiDAR point clouds as described in claim 1, characterized in that: Collect raw point cloud data and preprocess it to obtain clean surface point clouds. Construct a multi-scale digital elevation model sequence, including the following steps: Airborne LiDAR scanning was performed on the target slope to obtain raw point cloud data. The raw point cloud data was then denoised and filtered to obtain a clean surface point cloud. A multi-scale digital elevation model sequence is constructed based on clean surface point clouds through three-dimensional voxel downsampling. The multi-scale digital elevation model sequence includes digital elevation models at macro, meso, and meso scales.

3. The method for identifying potential slip surfaces of slopes based on multi-scale fusion of airborne LiDAR point clouds as described in claim 2, characterized in that: Extracting terrain features from a multi-scale digital elevation model sequence to obtain a multi-scale terrain feature map includes the following steps: Slope and aspect features are extracted from macro-scale digital elevation models in a multi-scale digital elevation model sequence using a grid neighborhood elevation difference algorithm. From the mesoscale digital elevation model in the multi-scale digital elevation model sequence, the plan curvature and profile curvature features are extracted using quadratic surface fitting. From the mesoscale digital elevation model in the multi-scale digital elevation model sequence, the surface roughness and normal vector variability features are extracted using the local surface fitting method of principal component analysis. Multi-scale topographic feature maps are obtained by fusing slope and aspect characteristics, planar curvature and profile curvature characteristics, and surface roughness and normal vector variability characteristics.

4. The method for identifying potential slip surfaces of slopes based on multi-scale fusion of airborne LiDAR point clouds as described in claim 3, characterized in that: The geometric boundaries of the mesoscale digital elevation model in the multi-scale digital elevation model sequence are combined with rainfall conditions set based on historical rainfall records to conduct unsaturated seepage simulation, obtain pore water pressure distribution cloud maps, and fuse multi-scale topographic feature maps to obtain the geometric slip sensitivity index, including the following steps: The triangular mesh surface of the spatial mesoscale digital elevation model in the multi-scale digital elevation model sequence is used as the geometric boundary input to the unsaturated seepage numerical solver. The unsaturated seepage numerical solver is loaded with rainfall conditions set based on historical rainfall records. The unsaturated seepage numerical solver is run to solve the Richards equation and obtain the pore water pressure distribution cloud map in the slope body. Normalize the slope features, aspect features, plan curvature features and profile curvature features, surface roughness features and normal vector variability features in multi-scale topographic feature maps; Adaptive weights are calculated based on the variances of normalized slope characteristics, aspect characteristics, plane curvature characteristics, profile curvature characteristics, surface roughness characteristics, and normal vector variability characteristics. The geometric slip sensitivity index is obtained by weighting and fusing the normalized slope features, aspect features, plane curvature features, profile curvature features, surface roughness features and normal vector variability features using adaptive weights.

5. The method for identifying potential slip surfaces of slopes based on multi-scale fusion of airborne LiDAR point clouds as described in claim 4, characterized in that: The critical slip surface index map is generated by coupling the pore water pressure distribution contour map with the geometric slip sensitivity index, including the following steps: The shear strength parameters obtained by direct shear test of soil and rock mass of target slope are combined with the critical threshold of pore water pressure value in pore water pressure distribution cloud map. The pore water pressure value at each location in pore water pressure distribution cloud map is compared with the critical threshold. The dynamic weight coefficient of each location is calculated based on the comparison results. The critical slip surface index map is generated by multiplying the dynamic weighting coefficients with the geometric slip sensitivity index values ​​in the geometric slip sensitivity index map.

6. The method for identifying potential slip surfaces of slopes based on multi-scale fusion of airborne LiDAR point clouds as described in claim 5, characterized in that: Thresholding and morphological processing are performed on the critical slip surface index map to identify and vectorize potential slip surface boundaries, including the following steps: The Otsu's method is used to perform threshold segmentation on the critical slip surface index map, and regions with index values ​​higher than the segmentation threshold are extracted to obtain a binary image. Morphological closing operations are performed on binary images to connect adjacent discrete regions in the binary image and fill the holes inside the regions; Edge tracing is performed on the morphologically processed connected regions to obtain their contour coordinates. The contour coordinates are converted into vector polygons to obtain the potential slip surface boundary.

7. The method for identifying potential slip surfaces of slopes based on multi-scale fusion of airborne LiDAR point clouds as described in claim 6, characterized in that: Based on the point error distribution of the original point cloud data and the permeability coefficient range of the soil and rock mass from the unsaturated seepage numerical simulation, the Monte Carlo simulation method is used to generate potential slip surface boundary samples with different spatial morphologies. These samples are then mapped to a three-dimensional spatial coordinate system defined by a multi-scale digital elevation model sequence for statistical analysis, generating a three-dimensional probability distribution field with confidence intervals. The process includes the following steps: A probability distribution model of point error is established based on the point error distribution of the original point cloud data, and a probability distribution model of permeability coefficient is established based on the permeability coefficient range of soil and rock mass in unsaturated seepage numerical simulation. The Monte Carlo simulation method was used to randomly select a set of parameter values ​​from the normal distribution model of the point error and the probability distribution model of the permeability coefficient, respectively. The original point cloud data is randomly perturbed using parameter values ​​to generate perturbed point cloud data. Based on the perturbed point cloud data, a multi-scale digital elevation model sequence is reconstructed using a three-dimensional voxelization downsampling method. Using parameter values ​​extracted from the probability distribution model of the permeability coefficient and combined with rainfall conditions set by historical rainfall records, a new set of pore water pressure distribution cloud maps is generated. Multi-scale topographic feature maps are extracted from the reconstructed multi-scale digital elevation model sequence, and the new pore water pressure distribution cloud map is fused with the multi-scale topographic feature maps to generate a new critical slip surface index map. Threshold segmentation and morphological processing are performed on the new critical slip surface index map to identify and vectorize a new potential slip surface boundary, and a set of potential slip surface boundary samples with different spatial morphologies is generated. Each potential slip surface boundary sample in the set of potential slip surface boundary samples with different spatial morphologies is uniformly mapped to the three-dimensional spatial coordinate system defined by the reconstructed multi-scale digital elevation model sequence. Spatial statistics are performed on the mapped potential slip surface boundary samples, and the kernel density estimation algorithm is used to calculate the frequency of sample occurrence at each spatial location. The frequency of sample occurrence is normalized to a probability value between zero and one, generating a three-dimensional probability distribution field with confidence intervals.

8. The method for identifying potential slip surfaces of slopes based on multi-scale fusion of airborne LiDAR point clouds as described in claim 7, characterized in that, Visualizing the three-dimensional probability distribution field with confidence intervals and outputting a slope stability analysis report includes the following steps: Import the 3D probability distribution field with confidence intervals into the 3D rendering engine; In the 3D rendering engine, a probabilistic rendering threshold is set, and the spatial geometric parameters of high-risk areas are extracted based on the rendering results. The spatial geometric parameters of high-risk areas are then integrated with geological survey information to generate a slope stability analysis report.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the slope potential slip surface identification method based on airborne LiDAR point cloud multi-scale fusion as described in any of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the slope potential slip surface identification method based on airborne LiDAR point cloud multi-scale fusion as described in any of claims 1 to 8.

Images