Sliding surface geometric inversion method coupling InSAR deformation data and numerical simulation

Abstract

The invention is suitable for the technical field of slip surface inversion, and provides a slip surface geometric inversion method coupling InSAR deformation data and numerical simulation, and the method comprises the steps: obtaining SAR image data, DEM data and drilling data of a landslide region; constructing a two-dimensional geologic model of the landslide area in three-dimensional simulation software based on the DEM data and the drilling data; optimizing the z coordinate vector by using a particle swarm optimization algorithm to obtain an optimal z coordinate vector which has a minimum fitness function and meets a preset spatial constraint; and constructing a sliding surface curve based on the optimal z coordinate vector. According to the invention, the efficiency and objectivity of sliding surface inversion can be improved.

Description

Technical Field

[0001] This application belongs to the field of slip surface inversion technology, and in particular relates to a slip surface geometric inversion method that couples InSAR deformation data with numerical simulation. Background Technology

[0002] The geometry of the landslide surface is a core physical parameter controlling the stability, kinematic characteristics, and disaster-causing range of a landslide. Accurate identification of the deep landslide surface is crucial in landslide risk assessment, monitoring and early warning, and prevention engineering design.

[0003] Currently, the main technical means for obtaining the geometry of landslide surfaces include:

[0004] Direct geological exploration methods, represented by borehole exploration, involve laying one or more boreholes on the landslide body to directly contact and identify the rock-soil interface, obtaining depth information of the slip surface at the location. Supplemented by engineering methods such as geological drilling and tunnel exploration, high-precision local slip surface locations can be obtained.

[0005] Geophysical exploration methods include high-density resistivity methods, ground-penetrating radar, and seismic exploration. These methods infer the spatial distribution of potential slip interfaces by detecting differences in physical properties (such as electrical properties, magnetism, and wave velocity) between the slip body and the bedrock.

[0006] Inversion method based on observation data and simplified model: The surface deformation field is obtained by using wide-area monitoring technology such as synthetic aperture radar interferometry (InSAR), and combined with the law of conservation of mass or simple rigid body motion assumptions, the approximate boundary and depth of the sliding surface are inferred by establishing the volume equilibrium relationship of the sliding body.

[0007] However, all of the above solutions have obvious inherent flaws:

[0008] Direct geological exploration methods are costly and have limited coverage: While methods such as drilling offer high accuracy at specific points, they are extremely expensive, have long construction periods, and have a certain environmental impact. For large or high-altitude landslides, inferences can only be made through a limited number of boreholes, making it difficult to obtain a complete landslide surface morphology, resulting in the limitation of "representing the entire surface with a single point."

[0009] Geophysical exploration methods suffer from multiple solutions and uncertainties: the interpretation of geophysical exploration results heavily relies on professional experience and is easily affected by various factors such as topography, groundwater, and heterogeneity of soil and rock masses, resulting in significant multiple solutions and uncertainties in the inversion results, making it difficult to guarantee accuracy.

[0010] Inversion methods based on observational data and simplified models are disconnected from physical reality: Methods relying on InSAR data and principles such as mass conservation typically simplify landslides into rigid bodies or simple motion patterns, ignoring the complex stress-strain relationships and nonlinear mechanical behavior within the landslide body. This simplification leads to unclear physical meaning of the inversion results and raises questions about their reliability.

[0011] In summary, existing landslide detection and inversion technologies have significant technical bottlenecks in terms of efficiency and objectivity. In particular, when facing high-altitude, remote landslides with complex terrain and inconvenient transportation, the limitations of traditional methods become increasingly prominent, making it difficult to meet the urgent needs of refined prevention and control of geological disasters. Summary of the Invention

[0012] This application provides a method for sliding surface geometric inversion that couples InSAR deformation data with numerical simulation, which can solve the problems of low efficiency and low objectivity in sliding surface inversion.

[0013] This application provides a method for geometric inversion of slip surfaces that couples InSAR deformation data with numerical simulation, including:

[0014] Acquire SAR image data, DEM data, and borehole data of the landslide area;

[0015] A two-dimensional geological model of the landslide area was constructed in 3D simulation software based on DEM data and borehole data.

[0016] The z-coordinate vector is optimized using the particle swarm optimization algorithm to obtain the optimal z-coordinate vector that minimizes the fitness function and satisfies the preset spatial constraints; the z-coordinate vector contains Elevation values ​​of each control point The control points are the midpoints on the slip surface curve, which is used to invert the slip surface of the landslide area. The starting point of the slip surface curve is... and the end point All control points are fixed at the landslide surface outcrop points in the landslide area. , Indicate control points The coordinate value in the x-axis direction, Indicate control points Elevation value, The x-axis represents the length of the landslide slope; the fitness function is calculated based on SAR image data and a two-dimensional geological model.

[0017] Construct a smooth curve based on the optimal z-coordinate vector.

[0018] Optionally, in the 3D simulation software, the surface of the 2D geological model is equipped with monitoring points that correspond one-to-one with the spatial locations of multiple InSAR observation points.

[0019] Optionally, in the process of optimizing the z-coordinate vector using the particle swarm optimization algorithm, the calculation process of the particle fitness function includes:

[0020] In the 3D simulation software, the slip surface corresponding to the particle is implanted into the 2D geological model, and the 3D simulation software is run to obtain the 3D displacement field of the ground surface at multiple monitoring points.

[0021] The three-dimensional displacement field of the Earth's surface at each monitoring point is projected onto the satellite's line of sight to obtain the simulated displacement value for each monitoring point;

[0022] The fitness function of particles is calculated based on SAR image data and simulated displacement values ​​from multiple monitoring points.

[0023] Optionally, based on SAR image data and simulated displacement values ​​from multiple monitoring points, the fitness function of the particles is calculated, including:

[0024] The fitness function of a particle is calculated using the following formula. :

[0025] ;

[0026] in, Indicates the core data item. This represents the coefficient of the preset convexity penalty term. This represents the convexity penalty term. This represents the coefficient of the preset curvature smoothing penalty term. This represents the curvature smoothing penalty term;

[0027] ;

[0028] ;

[0029] ;

[0030] in, Indicates the number of monitoring points. Indicates the first Simulated displacement values ​​at each monitoring point Indicates the first Displacement values ​​of each InSAR observation point It was extracted from the displacement field obtained after image processing of SAR image data. This indicates the number of sampling points used to uniformly sample the smooth surface curve. This indicates the first uniform sampling of the smooth surface curve. The coordinates of each sampling point along the x-axis. For the weight function, , Indicates the surface convexity points on the slip surface curve. This indicates the preset control parameters. Indicates the smooth surface curve at The second derivative of each sampling point.

[0031] Optionally, the three-dimensional displacement field of the Earth's surface at each monitoring point is projected onto the satellite line of sight to obtain the simulated displacement value for each monitoring point, including:

[0032] The number is calculated using the following formula. Simulated displacement values ​​at each monitoring point :

[0033] ;

[0034] in, Indicates the first The three-dimensional displacement field of the Earth's surface at each monitoring point This represents the unit vector in the LOS direction.

[0035] Optionally, the preset space constraints are:

[0036] ;

[0037] ;

[0038] in, Indicates the starting point The coordinate value in the x-axis direction, Indicates the end point The coordinate value in the x-axis direction, This indicates the lower limit of the preset burial depth of the sliding surface. Indicates the starting point At the elevation value, Indicates the end point Elevation value.

[0039] Optionally, a slip surface curve can be constructed based on the optimal z-coordinate vector, including:

[0040] Determining the smooth surface curve based on the optimal z-coordinate vector One control point;

[0041] Connect the starting points sequentially according to the x-axis coordinate values ​​in ascending order. , Control points and endpoints This yields the smooth surface curve.

[0042] The above-mentioned solution in this application has the following beneficial effects:

[0043] In the embodiments of this application, the z-coordinate vector is optimized based on high-precision SAR image data and a two-dimensional geological model using a particle swarm optimization algorithm to invert the optimal depth of the slip surface. Then, the slip surface curve of the landslide area is obtained based on the optimal depth. Compared with traditional methods, the slip surface inversion method of this application replaces a large number of subjective experience-dependent steps in traditional methods with an objective data-driven approach. Its automated and closed-loop inversion mechanism greatly improves efficiency and objectivity compared with traditional inversion methods that require manual intervention.

[0044] Other beneficial effects of this application will be described in detail in the following detailed description section. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 A flowchart of a method for geometric inversion of a slip surface coupled with InSAR deformation data and numerical simulation provided in an embodiment of this application;

[0047] Figure 2 Satellite image of a landslide in a village during the experiment;

[0048] Figure 3 This is a schematic diagram of the LOS-oriented deformation field of a landslide in a village during an experiment.

[0049] Figure 4 This is a schematic diagram of a landslide and its cross-section in a village during the experiment;

[0050] Figure 5 This is a schematic diagram of the slip surface inversion results for a certain village in the experiment. Detailed Implementation

[0051] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0052] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0053] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0054] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0055] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0056] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0057] To address the issues of low efficiency and objectivity in current slip surface inversion methods, this application provides a slip surface geometric inversion method that couples InSAR deformation data with numerical simulation. This method uses a particle swarm optimization algorithm to optimize the z-coordinate vector based on high-precision SAR image data and a two-dimensional geological model, inverting the optimal depth of the slip surface. Then, based on the optimal depth, the slip surface curve of the landslide area is obtained. Compared to traditional methods, this slip surface inversion method replaces a large number of subjective experience-dependent steps with an objective, data-driven approach. Its automated, closed-loop inversion mechanism significantly improves efficiency and objectivity compared to traditional inversion methods that require manual intervention.

[0058] The following describes the method for geometric inversion of slip surfaces by coupling InSAR deformation data with numerical simulation provided in this application, with reference to specific embodiments.

[0059] like Figure 1 As shown in the embodiments of this application, the method for geometric inversion of slip surfaces by coupling InSAR deformation data and numerical simulation includes the following steps:

[0060] Step 11: Obtain SAR image data, DEM data, and borehole data for the landslide area.

[0061] The aforementioned landslide area is the region requiring slip surface inversion. The aforementioned SAR image data can be obtained through satellite acquisition, for example, SAR image data of the landslide area acquired by the Sentinel-1 satellite. For instance, the aforementioned SAR image data could consist of 26 Sentinel-1A rising-orbit SAR images. After obtaining the SAR image data, small baseline set interferometry (SBAS-InSAR) can be used for temporal processing of the SAR images. This technique, through phase filtering, unwrapping, and atmospheric delay correction, can identify and jointly analyze temporary coherence points in densely vegetated, low-coherence areas, ultimately extracting a high spatial density, cumulative surface deformation field along the satellite line-of-sight, and then obtaining the displacement field based on the deformation field.

[0062] The aforementioned DEM data can be obtained by conducting on-site measurements of the landslide area using equipment such as total stations and Global Positioning System (GPS), or it can be generated using digital photogrammetry technology.

[0063] The aforementioned borehole data includes borehole spatial information (such as coordinates and depth), stratigraphic lithology classification, key interface depth (i.e., slip surface location), and geological exploration data such as physical and mechanical parameters obtained through in-situ testing and laboratory geotechnical tests. Specifically, core samples can be obtained through drilling to identify interfaces, borehole locations can be determined through surveying and positioning techniques, and material parameters can be measured through in-situ and laboratory geotechnical tests.

[0064] Step 12: Construct a two-dimensional geological model of the landslide area in 3D simulation software based on DEM data and borehole data.

[0065] In some embodiments of this application, the aforementioned three-dimensional simulation software can be the three-dimensional finite difference simulation software FLAC3D. Specifically, the DEM data and borehole data of the landslide area can be output to FLAC3D, and then a two-dimensional geological model (i.e., a two-dimensional profile benchmark geomechanical model) consistent with the real terrain can be constructed using FLAC3D. The model is assigned Mohr-Coulomb constitutive parameters and a groundwater level is set. Its boundary conditions are set as follows: the bottom is fixed (constraining displacement in the x and y directions), and the sides are normal constraints (constraining displacement in the x direction). Monitoring points corresponding to the spatial locations of InSAR observation points are set on the surface of the model (i.e., in the three-dimensional simulation software, the surface of the two-dimensional geological model is set with monitoring points corresponding one-to-one with the spatial locations of multiple InSAR observation points), and the model that has completed the initial stress equilibrium is saved as the unified initial state for all subsequent iterative calculations. Here, the x direction (i.e., the x-axis direction) is the length direction of the landslide area slope, and the y direction (i.e., the y-axis direction) is the width direction of the landslide area slope.

[0066] In some embodiments of this application, Akima spline curves are used to parametrically characterize the sliding surface morphology in order to achieve inversion. This curve is derived from... Control points Definition. Wherein, the termination point... and Fixed at the landslide surface outcrop point, in the middle The vertical (i.e., z-axis direction, also known as elevation direction) coordinates of each control point constitute the z-coordinate vector to be inverted. : , Indicate control points Elevation value, .

[0067] Step 13: Optimize the z-coordinate vector using the particle swarm optimization algorithm to obtain the optimal z-coordinate vector that minimizes the fitness function and satisfies the preset spatial constraints.

[0068] In the initial population of the particle swarm optimization algorithm described above, each particle is a z-coordinate vector, and all particles are distinct. It is understandable that the particles in the initial population can be randomly generated. The termination condition for the particle swarm optimization algorithm can be that the number of iterations reaches a preset number (e.g., 566), or the change in the minimum fitness function is less than a preset value, etc.

[0069] The above z-coordinate vector includes Elevation values ​​of each control point The control points are the midpoints on the slip surface curve (i.e., the aforementioned Akima spline curve). The slip surface curve is used to invert the slip surface of the landslide area, and the starting point of the slip surface curve is... and the end point All control points are fixed at the landslide surface outcrop points in the landslide area. , Indicate control points The coordinate value in the x-axis direction, Indicate control points Elevation value, The x-axis direction is the length direction of the slope surface in the landslide area.

[0070] To ensure the physical validity of the inversion results, the following two strict solution space constraints (i.e., the aforementioned pre-defined space constraints) are imposed on the z-coordinate vector:

[0071] (1) Sequence constraint: The x-coordinates of the control points must strictly increase from the trailing edge to the front edge to ensure that the sliding surface (i.e., the sliding surface) does not intersect itself.

[0072] ;

[0073] (2) Spatial boundary constraints: The depth of any control point must be between the ground surface and the preset minimum depth line.

[0074] ;

[0075] in, Indicates the starting point The coordinate value in the x-axis direction, Indicates the end point The coordinate value in the x-axis direction, This indicates the lower limit of the preset burial depth of the slip surface (i.e., the elevation value corresponding to the preset minimum depth line). Settings can be configured based on exploration data. Indicates the starting point At the elevation value, Indicates the end point Elevation value.

[0076] It should be noted here that the starting and ending points of the slip surface curve are easily observed due to the obvious bulges on the landslide surface. Therefore, the locations of significant surface deformation are designated as the starting and ending points. Consequently, the coordinates of the starting and ending points (i.e., the coordinates along the x-axis and z-axis) are known and can be obtained from exploration data (existing profiles). Furthermore, since this application uses control points for interpolation to generate the slip surface, the number and distribution of these control points can be set according to actual conditions, for example, setting five points and ensuring they are evenly distributed. Therefore, the x-axis coordinates of each control point are also known.

[0077] In some embodiments of this application, during the optimization of the z-coordinate vector using the particle swarm optimization algorithm, the particle fitness function is calculated based on SAR image data and a two-dimensional geological model. Specifically, the calculation process of the particle fitness function includes the following steps 13.1 to 13.3:

[0078] Step 13.1: In the 3D simulation software, the slip surface corresponding to the particle is implanted into the 2D geological model, and the 3D simulation software is run to obtain the 3D displacement field of the surface at multiple monitoring points.

[0079] In other words, in the 3D simulation software, the coordinates of each control point are determined based on the z-coordinate vector of the particles. Then, the slip surface curve is determined based on these coordinates. This slip surface curve is then embedded into a 2D geological model, and the 3D simulation software is run to simulate landslide behavior. After simulating landslide behavior in the 3D simulation software, the 3D surface displacement field of each monitoring point can be extracted from the software.

[0080] Step 13.2: Project the three-dimensional displacement field of the ground surface at each monitoring point onto the satellite line of sight to obtain the simulated displacement value of each monitoring point.

[0081] In some embodiments of this application, the first [number] is calculated using the following formula. Simulated displacement values ​​at each monitoring point :

[0082] ;

[0083] in, Indicates the first The three-dimensional displacement field of the Earth's surface at each monitoring point This represents the unit vector in the LOS direction.

[0084] Step 13.3: Calculate the fitness function of the particles based on SAR image data and simulated displacement values ​​from multiple monitoring points.

[0085] In some embodiments of this application, the fitness function of a particle can be calculated using the following formula. :

[0086] ;

[0087] in, Indicates the core data item. This represents the coefficient of the preset convexity penalty term. This represents the convexity penalty term. It is specifically designed to suppress unreasonable localized protrusions. This represents the coefficient of the preset curvature smoothing penalty term. This represents the curvature smoothing penalty term. Used to control the overall curvature of the curve and avoid sharp inflection points.

[0088] ;

[0089] ;

[0090] ;

[0091] in, Indicates the number of monitoring points. Indicates the first Simulated displacement values ​​at each monitoring point Indicates the first Displacement values ​​of each InSAR observation point It was extracted from the displacement field obtained after image processing of SAR image data. This indicates the number of sampling points used to uniformly sample the smooth surface curve. This indicates the first uniform sampling of the smooth surface curve. The coordinates of each sampling point along the x-axis. Here is the weight function, expressed as: , Indicates the surface convexity points on the slip surface curve. This indicates the preset control parameters. Indicates the smooth surface curve at The second derivative of each sampling point.

[0092] Indicates a specific location (a point of protrusion on the earth's surface). The function that adjusts the intensity of the convexity penalty, the control parameter σ determines the degree of this relaxation, and its value is taken with reference to the scale of the landslide and the morphology of the surface, so that the constraint can reflect the natural correspondence between the morphology of the slip surface and the surface profile.

[0093] fitness function To penalize unreasonable sliding surface shapes, this application minimizes the fitness function during the particle swarm optimization algorithm optimization process. With the goal of generating a new generation of candidate solutions, the particle swarm positions are updated according to the standard velocity-displacement calculation formula (using the calculation formula in the traditional particle swarm optimization algorithm) based on the historical best of individual particles and the global best of the swarm.

[0094] It should be noted that step 13 above is the core of the inversion loop. Through continuous simulation and comparison, it drives the solution to approach the optimal value until the optimization algorithm meets the convergence criterion.

[0095] Step 14: Construct the smooth surface curve based on the optimal z-coordinate vector.

[0096] In some embodiments of this application, the slip surface curve can be determined first based on the optimal z-coordinate vector. Control points Then, connect the starting points sequentially according to the x-axis coordinate values ​​in ascending order. , Control points and endpoints The slip surface curve is obtained, which is the result of the geometric inversion of the slip surface in the landslide area. It should be noted that since the slip surface inversion is displayed on a two-dimensional landslide profile, the y-axis direction is ignored, and the inversion result is a curve composed of xz axis coordinates.

[0097] It should be noted that the slip surface curve obtained by inversion in this application can be compared with the geological interface revealed by the borehole to prove the reliability and engineering practical value of the inversion method in this application.

[0098] Specifically, the optimal slip surface parameters (i.e., slip surface curves) obtained through inversion can be used to construct the final deterministic numerical model, and the results can be verified and visualized. The accuracy of the inversion results is quantitatively assessed by comparing the inverted slip surface depth with independent geological data not involved in the inversion (such as reserved verification borehole data). Finally, a comprehensive landslide profile map is generated, which visually displays the surface topography, the optimal slip surface line obtained through inversion, the optimized control point locations, and the spatial distribution of materials in the landslide body and bedrock, providing a clear geometric basis for subsequent engineering applications.

[0099] The core technological advantage of this application lies in the deep coupling of high-precision remote sensing data with physical and mechanical models. It replaces the traditional methods that rely heavily on subjective experience with an objective data-driven approach. Its automated and closed-loop inversion mechanism greatly improves efficiency and objectivity compared to manual trial and error. The final output of the slip surface results not only matches the surface deformation geometrically, but is also self-consistent in terms of physical and mechanical aspects, providing a highly reliable model foundation for subsequent accurate landslide stability evaluation and disaster risk assessment.

[0100] The feasibility of the inversion method of this application will be illustrated by specific experiments below.

[0101] In this experiment, a landslide in a village in 2024 was used as the research object. The study area was a town in a county of a city in a province, with a total area of ​​approximately 6,832 square meters. The terrain of the county is high in the southeast and northeast and low in the southwest. The landform type of the study area is erosion-densification hilly landform. The main sliding direction of the landslide was 298° (see...). Figure 2 , Figure 2The landslide, shaped like a long strip (resembling a foot), exhibits a steep front edge, a gentler middle and rear section, and a steep rear edge in its longitudinal profile. After the landslide, the surrounding terrain is higher than the landslide body, indicating significant subsidence and a clear landslide boundary. The shear outlet elevation at the front is 151-153 m, and the rear elevation is 275-280 m, a relative height difference of 125 m. The horizontal length from the shear outlet to the rear edge of the landslide body is approximately 400 m, the slope length is approximately 560 m, and the slope angle is 20°–30°. The length from the rear edge of the landslide body to the front of the landslide tongue in the accumulated material is approximately 490 m, and the width of the landslide tongue is approximately 350 m. Given the scale and complex morphological characteristics of this landslide, a detailed post-disaster investigation is particularly urgent. Since the morphology of the slip surface directly controls the residual stability of the slip body and its future development trend (such as creep or reactivation instability), and provides the most fundamental physical basis for subsequent risk assessment and the formulation of scientific disaster prevention and mitigation strategies, accurately determining the spatial location and geometric morphology of the deep main slip surface becomes the core link in various exploration tasks.

[0102] In this study, 26 Sentinel-1 A images within the study area were collected as experimental data. The specific parameters of the InSAR data are shown in Table 1. The relevant data can be downloaded from the official website of the EU Copernicus data ecosystem (https: / / dataspace.copernicus.eu / ).

[0103] Table 1. Statistics of InSAR Data Parameters

[0104] Considering that SBAS-InSAR technology can effectively overcome spatiotemporal incoherence and suppress topographic and atmospheric errors, and is currently widely used in temporal deformation monitoring of geological disasters such as earthquakes and landslides, this experiment uses SBAS-InSAR technology to process image data of the study area, and obtains relevant deformation information such as... Figure 3 As shown. Figure 3 In this context, longitude represents longitude, latitude represents latitude, and Cumulative LOS Deformation represents cumulative LOS deformation.

[0105] A schematic diagram of the existing geological exploration lines in the landslide area is shown below. Figure 4 As shown, this experiment uses profile 2-2' for analysis. The hardware configuration is as follows: Intel Core i7-13700 processor (2.10 GHz, 16 cores and 24 threads), 16 GB of memory; the software environment is FLAC3D 7.0 numerical simulation software. The experiment uses the particle swarm optimization (PSO) algorithm, and the specific parameter settings are shown in Table 2. Figure 4 In the diagram, ZK2 represents borehole 2, ZK7 represents borehole 7, ZK11 represents borehole 11, ZK15 represents borehole 15, ZK19 represents borehole 19, and 1-1', 3-3', 4-4', and 5-5' all represent cross-sections.

[0106] Table 2 PSO Parameter Settings Table

[0107] Experimental results are as follows Figure 5 As shown, Figure 5 In the table, Inverted Slip Surface and Geotechnical Model represents the geotechnical engineering model; Distance represents distance; Elevation represents elevation; Ground Surface represents the surface line; Water Table represents the groundwater level; Existing slip surface represents the existing (prior) slip surface; Inverted Slip Surface represents the inverted slip surface; Optimized Control Points represent the optimized control points; Unsaturated Slide Material represents the unsaturated slip material; Saturated Slide Material represents the saturated slip material; Bedrock represents the bedrock; and Monitoring Points (InSAR) represent the monitoring points.

[0108] The optimal slip surface (red solid line) obtained by this experiment shows a high degree of consistency with the slip surface (orange solid line) given by existing exploration data. The root mean square error (RMSE) of the simulated surface deformation and the actual deformation data observed by SBAS-InSAR converges to 0.0079 meters. A geometric comparison was performed between the inverted slip surface and the "existing slip surface" assumed by prior knowledge. The comparison results show that the absolute error (MAE) of the two slip surfaces is 3.9857 meters, and the average relative deviation is only 3.16%. These results strongly demonstrate the effectiveness of this application. Furthermore, the small deviation scientifically reflects that this invention uses continuous InSAR observation data to objectively and accurately optimize and refine the model obtained by traditional discrete exploration methods, thus obtaining a slip surface morphology closer to physical reality, providing a more reliable geometric model for subsequent stability analysis. In summary, the experiments show that the inversion method of this application can objectively and accurately invert the internal slip surface morphology of complex landslides, fully meeting the technical requirements for refined geological hazard exploration and risk assessment.

[0109] In summary, the inversion method of this application has the following advantages:

[0110] Significantly improved inversion accuracy and physical realism: This application uses geotechnical mechanics numerical simulation as the physics engine, ensuring that the inversion process is always strictly constrained by real physical laws such as stress-strain relationships. Compared to simplified models that rely solely on mass conservation or rigid body assumptions, the inversion results of this application are improved from "kinematically possible" to "mechanically reasonable." The output slip surface morphology can realistically reflect the mechanical response inside the slip surface and can be directly used for subsequent stability calculations, fundamentally improving the reliability and accuracy of engineering applications.

[0111] The efficiency and automation of the inversion process are significantly improved: Traditional manual trial-and-error methods relying on numerical simulations typically require a senior engineer to spend several working days repeatedly adjusting slip surface parameters, running the model, and comparing the results. This application constructs a closed-loop automatic interaction framework between the PSO algorithm and FLAC3D software, which can systematically complete thousands of iterations of optimization in a shorter unattended computing time, reducing the effective manual intervention time by more than 95%, greatly improving the efficiency of slip surface identification, and meeting the timeliness requirements of disaster emergency assessment.

[0112] The objectivity and reproducibility of the results are significantly enhanced: This application replaces the subjective judgment that heavily relies on engineers' geological experience and intuition with a data-driven intelligent optimization algorithm. The search process for the slip surface is entirely dominated by observational data and physical models, fundamentally eliminating the subjective bias and randomness that varies from person to person in manual trial and error methods, making the inversion results highly objective and fully reproducible.

[0113] Model applicability and geometric description capability expansion: Compared with traditional methods that often simplify the slip surface to a circular arc or a simple broken line, this application uses Akima spline curves with high degrees of freedom to parameterize the slip surface, which can finely characterize the irregular, wavy and other complex slip surface morphologies formed under complex geological conditions, significantly broadening the application range and inversion accuracy of the technology in various complex landslides.

[0114] The above description is the preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principles described in this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for geometric inversion of slip surfaces coupled with InSAR deformation data and numerical simulation, characterized in that, include: Acquire SAR image data, DEM data, and borehole data of the landslide area; A two-dimensional geological model of the landslide area is constructed in three-dimensional simulation software based on the DEM data and the borehole data. The z-coordinate vector is optimized using a particle swarm optimization algorithm to obtain an optimal z-coordinate vector that minimizes the fitness function and satisfies preset spatial constraints; the z-coordinate vector includes Elevation values ​​of each control point The control point is the midpoint of the slip surface curve, which is used to invert the slip surface of the landslide area. The starting point of the slip surface curve is... and the end point All control points are fixed at the landslide surface outcrop points in the landslide area. , Indicate control points The coordinate value in the x-axis direction, Indicate control points Elevation value, The x-axis direction represents the length of the slope surface in the landslide area; the fitness function is calculated based on the SAR image data and the two-dimensional geological model. The smooth surface curve is constructed based on the optimal z-coordinate vector.

2. The method for geometric inversion of slip surfaces according to claim 1, characterized in that, In the three-dimensional simulation software, the surface of the two-dimensional geological model is equipped with monitoring points that correspond one-to-one with the spatial locations of multiple InSAR observation points.

3. The method for geometric inversion of slip surfaces according to claim 2, characterized in that, In the process of optimizing the z-coordinate vector using the particle swarm optimization algorithm, the calculation process of the particle fitness function includes: In the three-dimensional simulation software, the slip surface corresponding to the particle is implanted into the two-dimensional geological model, and the three-dimensional simulation software is run to obtain the three-dimensional displacement field of the ground surface at multiple monitoring points; The three-dimensional displacement field of the Earth's surface at each monitoring point is projected onto the satellite's line of sight to obtain the simulated displacement value for each monitoring point; The fitness function of the particles is calculated based on the SAR image data and the simulated displacement values ​​of multiple monitoring points.

4. The method for geometric inversion of slip surfaces according to claim 3, characterized in that, The calculation of the particle fitness function based on the SAR image data and simulated displacement values ​​from multiple monitoring points includes: The fitness function of a particle is calculated using the following formula. : ； in, Indicates the core data item. This represents the coefficient of the preset convexity penalty term. This represents the convexity penalty term. This represents the coefficient of the preset curvature smoothing penalty term. This represents the curvature smoothing penalty term; ； ； ； in, Indicates the number of monitoring points. Indicates the first Simulated displacement values ​​at each monitoring point Indicates the first Displacement values ​​of each InSAR observation point It is extracted from the displacement field obtained after image processing of the SAR image data. This indicates the number of sampling points used to uniformly sample the smooth surface curve. This indicates the first uniform sampling of the smooth surface curve. The coordinates of each sampling point along the x-axis. For the weight function, , Indicates the surface convexity points on the slip surface curve. This indicates the preset control parameters. Indicates the smooth surface curve at The second derivative of each sampling point.

5. The method for geometric inversion of slip surfaces according to claim 3, characterized in that, The process of projecting the three-dimensional displacement field of the Earth's surface at each monitoring point onto the satellite's line-of-sight direction to obtain the simulated displacement value for each monitoring point includes: The number is calculated using the following formula. Simulated displacement values ​​at each monitoring point : ； in, Indicates the first The three-dimensional displacement field of the Earth's surface at each monitoring point This represents the unit vector in the LOS direction.

6. The method for geometric inversion of slip surfaces according to claim 1, characterized in that, The preset spatial constraint is: ； ； in, Indicates the starting point The coordinate value in the x-axis direction, Indicates the end point The coordinate value in the x-axis direction, This indicates the lower limit of the preset burial depth of the sliding surface. Indicates the starting point At the elevation value, Indicates the end point Elevation value.

7. The method for geometric inversion of slip surfaces according to claim 6, characterized in that, The construction of the smooth surface curve based on the optimal z-coordinate vector includes: The slip surface curve is determined based on the optimal z-coordinate vector. One control point; Connect the starting points sequentially according to the x-axis coordinate values ​​in ascending order. , Control points and endpoints This yields the smooth surface curve.

Images