Search

CN-121999370-A - High-level landslide identification method by cooperation of multi-source remote sensing and ground investigation

CN121999370ACN 121999370 ACN121999370 ACN 121999370ACN-121999370-A

Abstract

The application discloses a high-level landslide identification method by cooperation of multi-source remote sensing and ground investigation, and belongs to the technical field of geological disaster monitoring and identification. The method comprises the steps of constructing a multisource information basic data set, generating deformation characteristic data representing deformation activity states, constructing environment characteristic data representing landslide formation conditions, screening landslide candidate areas meeting deformation activity criteria and landslide formation condition criteria simultaneously, conducting comprehensive characteristic analysis and type discrimination on the landslide candidate areas to obtain landslide identification conclusion and corresponding uncertainty information, determining a ground investigation and check scheme according to the uncertainty information and executing the landslide identification conclusion, taking the landslide identification conclusion as a final result and outputting corresponding landslide type and uncertainty information if the landslide identification conclusion is consistent, and correcting the landslide identification conclusion according to the check result and updating the uncertainty information, taking the corrected conclusion as a final result and outputting the corresponding landslide type and the updated uncertainty information if the landslide identification conclusion is inconsistent.

Inventors

  • XIE HANG
  • XIANG QIWEN
  • WANG JUAN
  • YANG JUAN
  • Yue Fazheng

Assignees

  • 贵州省第一测绘院(贵州省北斗导航位置服务中心)

Dates

Publication Date
20260508
Application Date
20260407

Claims (10)

  1. 1. A multi-source remote sensing and ground investigation collaborative high-level landslide identification method is characterized by comprising the following steps: Acquiring remote sensing observation data and ground investigation information of a target area, and constructing a multisource information base data set for high-order landslide identification; extracting surface deformation characteristic information based on the multisource information basic data set, identifying potential deformation abnormal areas, and generating deformation characteristic data representing deformation activity states; extracting topography feature, geological condition feature and human activity feature based on the multisource information basic data set, and constructing environment feature data representing landslide formation conditions; Coupling analysis is carried out on the deformation characteristic data and the environment characteristic data, and landslide candidate areas meeting deformation activity criteria and landslide formation condition criteria at the same time are screened; Carrying out comprehensive feature analysis and type discrimination on the landslide candidate area to obtain a landslide identification conclusion and corresponding uncertainty information; determining a ground investigation checking scheme according to the uncertainty information and executing the ground investigation checking scheme; if the check result is consistent with the landslide identification conclusion, taking the landslide identification conclusion as a final result, and outputting corresponding landslide type and uncertainty information; If the check result is inconsistent with the landslide identification conclusion, correcting the landslide identification conclusion according to the check result, updating the uncertainty information, taking the corrected conclusion as a final result, and outputting the corresponding landslide type and the updated uncertainty information.
  2. 2. The method of claim 1, wherein the remote sensing observation data comprises multi-phase satellite-borne synthetic aperture radar image data, high resolution optical image data, and airborne LiDAR point cloud data.
  3. 3. The high-order landslide recognition method of claim 2, wherein the method of acquiring the ground investigation information of the target area is: aiming at the potential geological disaster hidden trouble points which are already defined in the target area, a staged investigation scheme is formulated according to an air-space-ground integrated investigation principle, and investigation ranges, investigation contents and investigation modes of each stage are defined; Performing on-plane preliminary investigation by depending on regional geological data, and determining on-site check points by combining remote sensing interpretation results with existing geological structure information; Performing key verification on the determined site verification point; and collecting and arranging site investigation data in the key checking process to obtain the ground investigation information of the target area.
  4. 4. The high-order landslide identification method of claim 3, wherein the method for extracting the surface deformation characteristic information based on the multi-source information basic data set comprises the following steps: Constructing an interference processing data set based on the acquired multi-time-phase satellite-borne synthetic aperture radar image data, wherein the interference processing data set comprises a double-time-phase interference image pair and a long-time-sequence interference image sequence; Carrying out differential interferometry processing based on the double-phase interference image pair to generate a differential interference image, and sequentially carrying out terrain phase removal and atmospheric delay correction to obtain accumulated deformation distribution data of the sight direction of the target area; Performing interferometry processing of a permanent scatterer based on the long-time-sequence interference image sequence to obtain earth surface deformation rate data and earth surface displacement time sequence data of a high coherence point; Extracting a deformation rate section in a potential candidate area along a main sliding direction based on the surface deformation rate data and the surface displacement time sequence data, selecting representative monitoring points to construct a displacement-time curve, and judging a deformation stage where the surface deformation is located; and (3) arranging and storing the earth surface deformation rate data and the earth surface displacement time sequence data to obtain an earth surface deformation characteristic information data set of the target area.
  5. 5. The method for identifying a high-level landslide of claim 4, wherein the method for identifying the potential deformation abnormal region is as follows: Extracting a high deformation region according to a preset deformation rate threshold value based on the obtained earth surface deformation rate data, and dividing a plurality of candidate deformation connected regions according to space connectivity; Selecting representative monitoring points of the trailing edge, the middle part and the leading edge along the main sliding direction in each candidate deformation communication area, extracting annual average surface deformation rate of each monitoring point, calculating the ratio of the deformation rate difference value of the trailing edge to the leading edge to the corresponding space distance, and obtaining the deformation rate gradient characteristics from the trailing edge to the leading edge; counting the space distribution density and the linking degree of high-coherence points in each candidate deformation communication area to obtain deformation space continuity characteristics; Extracting surface displacement time sequence data of representative monitoring points in each candidate deformation communication area, constructing an accumulated displacement-time curve, analyzing the change trend of the curve, and obtaining time sequence deformation rule characteristics; Comprehensively analyzing the deformation rate gradient characteristics, the deformation space continuity characteristics and the time-sequence deformation rule characteristics and the earth surface deformation rate data to obtain deformation characteristic data, and identifying potential deformation abnormal areas according to the deformation characteristic data.
  6. 6. The high-order landslide recognition method of claim 4, wherein extracting the topographical features, the geological condition features and the human activity features based on the multi-source information base data set to construct the environmental feature data characterizing the landslide formation condition comprises the steps of: Based on the multi-source information basic data set, extracting digital elevation model data, high-resolution optical image data, regional geological data and field investigation data, and constructing an environment characteristic extraction data set; Based on the digital elevation model data and combined with high-resolution optical image data, gradient and slope calculation, topography relief and relative height difference calculation, slope micro-topography identification and catchment area extraction are sequentially carried out, gradient slope characteristic parameters, relative height difference characteristic parameters, slope micro-topography characteristic parameters and catchment area characteristic parameters are obtained, and the characteristic parameters are integrated to obtain topography and topography characteristic data; extracting lithology characteristic parameters, geological structure characteristic parameters and structural surface characteristic parameters based on regional geological data, and acquiring disaster distribution characteristic parameters by combining historical disaster data to integrate so as to obtain geological condition characteristic data; Identifying various ergonomic activities based on the high-resolution optical image data and the field investigation data, and calculating disturbance intensity indexes according to the influence degree of the various ergonomic activities on the slope stability to obtain the characteristic data of the ergonomic activities; And integrating the topographic feature data, the geological condition feature data and the human activity feature data to obtain the environmental feature data for representing landslide formation conditions.
  7. 7. The high-order landslide recognition method of claim 6, wherein the coupling analysis is performed on the deformation characteristic data and the environmental characteristic data, and the landslide candidate region meeting both the deformation activity criterion and the landslide formation condition criterion is screened out, comprising the steps of: Identifying deformation abnormal areas based on the earth surface deformation rate data, converting the deformation abnormal areas into planar space analysis units, and establishing attribute fields for the planar space analysis units; loading corresponding deformation characteristic data in a planar space analysis unit; Taking a planar space analysis unit as a statistical range, and extracting environmental characteristic data overlapped with the planar space analysis unit in space; Carrying out joint discrimination on the deformation characteristic data and the environment characteristic data, and identifying a deformation abnormal region meeting landslide formation conditions according to a preset discrimination rule; and determining the identified deformation abnormal region which meets the landslide formation condition and has the continuous deformation activity characteristic as a landslide candidate region.
  8. 8. The high-order landslide recognition method according to any one of claims 1 to 7, wherein the comprehensive feature analysis and type discrimination are performed on the landslide candidate region to obtain a landslide recognition conclusion and corresponding uncertainty information, and the method comprises the following steps: For a landslide candidate region, extracting gradient characteristics, slope body structural characteristics, lithology conditions, construction control characteristics, groundwater characteristics and ergonomic activity characteristics of the landslide candidate region, and constructing a comprehensive characteristic data set for distinguishing landslide types; based on a preset landslide type discrimination criterion system, carrying out matching analysis on the comprehensive characteristic data set, and determining disaster types corresponding to each landslide candidate area; extracting space morphological parameters of the landslide candidate region for completing type discrimination, and estimating length, width, thickness and volume parameters of a disaster body to form landslide identification result data; Based on the data coverage integrity, the observation condition sufficiency, the data source consistency and the ground checking degree, carrying out reliability quantitative evaluation on landslide identification result data to obtain a reliability quantitative evaluation result; and dividing and identifying confidence level for each landslide candidate area according to the reliability quantitative evaluation result, and integrating disaster type, scale parameter and confidence level to obtain landslide identification conclusion and corresponding uncertainty information.
  9. 9. The high-order landslide identification method of claim 8, wherein the method for determining the ground investigation and verification scheme based on the uncertainty information comprises the steps of: acquiring the identification confidence level corresponding to each landslide candidate region and the uncertainty information in the identification conclusion; Dividing the ground investigation checking priority of each landslide candidate area according to the recognition confidence level; determining a survey mode according to the topography condition and the accessibility of the landslide candidate area; performing on-site check aiming at an uncertainty source marked in the identification conclusion to acquire investigation data for verifying the landslide identification conclusion; and (5) comprehensive investigation and check priority, investigation mode and uncertainty information check results to obtain a ground investigation and check scheme.
  10. 10. The high-order landslide identification method of claim 9, wherein the method for correcting the landslide identification result according to the checking result comprises the following steps: obtaining ground investigation and check results corresponding to each landslide candidate area, and extracting boundary range information, deformation sign information and disaster type discrimination information related to landslide identification conclusion; Consistency comparison is carried out on the ground investigation and check result and the original landslide identification conclusion, and landslide candidate areas with identification differences are determined; Analyzing the reasons for identifying inconsistencies aiming at landslide candidate areas with identification differences, and determining corresponding inconsistent types; According to the inconsistent type, correcting the landslide identification conclusion, including redefining a landslide boundary, updating basic terrain data, supplementing remote sensing observation data or eliminating a misjudgment area; Recording landslide identification conclusions and correction basis before and after correction, and obtaining the landslide identification conclusions after correction by ground investigation.

Description

High-level landslide identification method by cooperation of multi-source remote sensing and ground investigation Technical Field The application relates to the technical field of geological disaster monitoring and recognition, in particular to a high-level landslide recognition method by cooperation of multi-source remote sensing and ground investigation. Background The high-level landslide is a geological disaster which occurs on a steep mountain or a significant slope with a high difference, and generally has the characteristics of strong concealment, high burstiness, wide damage range and the like, and once the landslide is unstable, serious threat is often caused to lower residential points, traffic facilities and engineering construction. Along with the continuous increase of mountain traffic construction, mineral development and engineering activities, the disaster-tolerant conditions of high-level landslide are increasingly complex, and timely and accurately identifying and distinguishing the disaster-tolerant conditions become important contents in geological disaster prevention and control work. Traditional landslide identification mainly relies on manual ground investigation and experience interpretation, and landslide bodies and deformation signs thereof are identified through on-site stepping investigation, geological mapping and other modes. Although the method can obtain more reliable engineering geological information, the method is often limited by terrain conditions in a high-level steep slope region, has poor accessibility and low investigation efficiency, and is difficult to develop system identification in a large-range region. Along with the development of remote sensing technology, multi-source remote sensing data are gradually applied to the field of geological disaster monitoring and identification. For example, the optical remote sensing image is used for performing morphological interpretation, the synthetic aperture radar interferometry technology is used for obtaining the surface deformation information, and the laser radar data is used for extracting high-precision topographic features and the like. The technology can acquire the surface morphology and deformation information in a large-scale area, and provides important data support for landslide identification. However, it is often difficult for single remote sensing data to comprehensively reflect the formation conditions and evolution characteristics of landslide, and the single remote sensing data is easily affected by factors such as vegetation coverage, observation conditions, terrain shielding and the like, so that erroneous judgment or missed judgment is generated. Therefore, how to comprehensively utilize multi-source remote sensing observation data and ground investigation information to perform collaborative analysis on the surface deformation characteristics and landslide formation environmental conditions and improve the accuracy and reliability of high-order landslide identification results has become a technical problem to be solved. Disclosure of Invention In order to overcome a series of defects in the prior art, the application aims to provide a high-level landslide identification method by cooperating multi-source remote sensing with ground investigation, which comprises the following steps: Acquiring remote sensing observation data and ground investigation information of a target area, and constructing a multisource information base data set for high-order landslide identification; extracting surface deformation characteristic information based on the multisource information basic data set, identifying potential deformation abnormal areas, and generating deformation characteristic data representing deformation activity states; extracting topography feature, geological condition feature and human activity feature based on the multisource information basic data set, and constructing environment feature data representing landslide formation conditions; Coupling analysis is carried out on the deformation characteristic data and the environment characteristic data, and landslide candidate areas meeting deformation activity criteria and landslide formation condition criteria at the same time are screened; Carrying out comprehensive feature analysis and type discrimination on the landslide candidate area to obtain a landslide identification conclusion and corresponding uncertainty information; determining a ground investigation checking scheme according to the uncertainty information and executing the ground investigation checking scheme; if the check result is consistent with the landslide identification conclusion, taking the landslide identification conclusion as a final result, and outputting corresponding landslide type and uncertainty information; If the check result is inconsistent with the landslide identification conclusion, correcting the landslide identification conclusion according to the check result,