Canyon reservoir landslide identification method, device, equipment and storage medium

Abstract

The application relates to the technical field of water disaster monitoring, in particular to a canyon-type reservoir area landslide identification method, device and equipment and a storage medium, wherein the method comprises the following steps: acquiring radar data and topography of a canyon-type reservoir area; determining target landslide deformation points of the canyon-type reservoir area according to the radar data and the topography, and calculating the deformation rate of the target landslide deformation points; generating the spatial aggregation relationship of the target landslide deformation points based on the deformation rate, and identifying a landslide area in the canyon-type reservoir area based on the deformation rate and the spatial aggregation relationship. Thus, the problems, such as low landslide area identification accuracy, of the prior art are solved.

Description

Technical Field

[0001] This application relates to the field of water conservancy disaster monitoring technology, and in particular to a method, device, equipment and storage medium for identifying landslides in canyon-type reservoir areas. Background Technology

[0002] Landslides refer to the phenomenon of large amounts of rock, debris, or soil moving down a slope. my country has a wide variety of geological hazards, and landslides are among the most destructive. Their main causes include rainfall, earthquakes, human activities, landscape alteration, and the natural process of slope erosion. However, most reservoirs are located in mountainous and canyon areas with complex geological structures, undulating terrain, numerous gullies, and steep slopes. Landslides along the reservoir banks are frequent, causing significant socio-economic losses and ecological damage. Therefore, identifying potential landslide hazards in water conservancy project areas and conducting landslide susceptibility assessments based on this identification is of great significance for regional water conservancy project planning and construction, reservoir geological hazard prevention, and is also a fundamental and important component of reservoir dam safety evaluation.

[0003] To mitigate and avoid the impact of landslides, the key is to accurately identify and detect potential landslides and take preventative measures in advance, thereby reducing the probability of landslides occurring. Traditional landslide monitoring mainly relies on GPS, total stations, and displacement gauges to monitor landslide movement, but this method has drawbacks such as high investment costs and limited monitoring range.

[0004] Synthetic Aperture Radar Interferometry (SAR) technology has attracted much attention due to its advantages such as all-weather, all-time, and large-scale surface deformation monitoring. It has become a new means for identifying geological hazards such as landslides and monitoring deformation. Among them, D-InSAR technology shows low correlation and decoherence of SAR images from different time periods; PS-InSAR technology uses a single image as a common master image, and decoherence will occur for image pairs with large interferometric time distances. Therefore, it is not suitable for deformation monitoring with large time spans, which leads to low accuracy in identifying landslide areas. Summary of the Invention

[0005] This application provides a method, apparatus, equipment, and storage medium for identifying landslides in canyon-type reservoir areas, in order to solve the problem of low accuracy in landslide area identification in existing technologies.

[0006] The first aspect of this application provides a method for identifying landslides in a canyon-type reservoir area, comprising the following steps: acquiring radar data and topography of the canyon-type reservoir area; determining target landslide deformation points in the canyon-type reservoir area based on the radar data and topography, and calculating the deformation rate of the target landslide deformation points; generating spatial clustering relationships of the target landslide deformation points based on the deformation rate, and identifying landslide areas in the canyon-type reservoir area based on the deformation rate and spatial clustering relationships.

[0007] Optionally, after identifying landslide areas in canyon-type reservoir areas based on deformation rate and spatial clustering relationships, the method further includes: assessing the landslide susceptibility level of the landslide areas; and issuing early warnings or alerts based on the landslide susceptibility level.

[0008] Optionally, calculating the deformation rate of the target landslide deformation point includes: obtaining the deformation rate along the line of sight corresponding to the target deformation point; and converting the deformation rate along the line of sight into a deformation rate perpendicular to the ground based on radar geometry and topography.

[0009] Optionally, the target landslide deformation points in the canyon-type reservoir area are determined based on radar data and topography, including: processing radar data to obtain multiple landslide deformation points and corresponding deformation rates in the canyon-type reservoir area; and filtering multiple landslide deformation points based on topography and deformation rates to obtain the target landslide deformation points in the canyon-type reservoir area.

[0010] Optionally, generating the spatial clustering relationship of the target landslide deformation points based on the deformation rate includes: calculating the spatial distribution pattern of the target landslide deformation points based on the deformation rate; and generating the spatial clustering relationship of the target landslide deformation points based on the spatial distribution pattern.

[0011] Optionally, assessing the landslide susceptibility level of a landslide area includes: obtaining at least one landslide influencing factor for the landslide area; calculating the normalized information value and weight of each landslide influencing factor; and assessing the landslide susceptibility level of the landslide area based on the normalized information value and weight.

[0012] Optionally, the normalized information value and weight of each landslide impact factor are calculated, including: classifying each landslide impact factor into levels and using an information value model to calculate the information value of each level of each landslide impact factor; unifying the information values ​​of each level of each landslide impact factor to obtain a normalized information value; using the classification characteristics of the landslide area and each landslide impact factor as samples; and performing a binary classification of the samples as landslide and non-landslide to obtain the weight of each landslide impact factor.

[0013] A second aspect of this application provides a landslide identification device for a canyon-type reservoir area, comprising: an acquisition module for acquiring radar data and topography of the canyon-type reservoir area; a determination module for determining target landslide deformation points in the canyon-type reservoir area based on the radar data and topography, and calculating the deformation rate of the target landslide deformation points; and an identification module for generating spatial clustering relationships of the target landslide deformation points based on the deformation rate, and identifying landslide areas in the canyon-type reservoir area based on the deformation rate and spatial clustering relationships.

[0014] Optionally, it also includes: an assessment module for assessing the landslide susceptibility level of landslide areas after identifying landslide areas in canyon-type reservoir areas based on deformation rate and spatial clustering relationships; and issuing early warnings or alerts based on the landslide susceptibility level.

[0015] Optionally, the determining module is further configured to: obtain the deformation rate in the line-of-sight direction corresponding to the target deformation point; and convert the deformation rate in the line-of-sight direction into a deformation rate perpendicular to the ground based on the radar geometry and terrain.

[0016] Optionally, the determining module is further used to: process radar data to determine multiple landslide deformation points and corresponding deformation rates in the canyon-type reservoir area; and filter multiple landslide deformation points based on topography and deformation rate to obtain the target landslide deformation point in the canyon-type reservoir area.

[0017] Optionally, the identification module is further used to: calculate the spatial distribution pattern of the deformation points of the target landslide based on the deformation rate; and generate the spatial clustering relationship of the deformation points of the target landslide based on the spatial distribution pattern.

[0018] Optionally, the assessment module is further used to: obtain at least one landslide influencing factor for the landslide area; calculate the normalized information value and weight of each landslide influencing factor; and assess the landslide susceptibility level of the landslide area based on the normalized information value and weight.

[0019] Optionally, the evaluation module is further used to: classify each landslide impact factor into levels and calculate the information value of each level of each landslide impact factor using an information value model; unify the information values ​​of each level of each landslide impact factor to obtain a normalized information value; use the classification characteristics of the landslide area and each landslide impact factor as samples; and perform landslide and non-landslide binary classification on the samples to obtain the weight of each landslide impact factor.

[0020] A third aspect of this application provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the program to perform the landslide identification method for canyon-type reservoir areas as described in the above embodiments.

[0021] The fourth aspect of this application provides a computer-readable storage medium having a computer program or instructions stored thereon, which is executed by a processor to perform the landslide identification method for canyon-type reservoir areas as described above.

[0022] Therefore, this application has at least the following beneficial effects:

[0023] This application's embodiments can determine target deformation points in canyon-type reservoir areas based on radar data and topographic features, calculate the deformation rate of these target landslide deformation points, and then generate spatial clustering relationships based on these deformation rates. By accurately identifying landslide areas within canyon-type reservoir areas based on both deformation rates and spatial clustering relationships, the accuracy of landslide area identification is improved, providing a foundation for subsequent landslide disaster early warning. This solves the technical problem of low accuracy in landslide area identification in existing technologies.

[0024] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0025] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0026] Figure 1 This is a flowchart of a landslide identification method for canyon-type reservoir areas provided according to an embodiment of this application;

[0027] Figure 2 This is a flowchart of landslide identification and susceptibility assessment provided according to embodiments of this application;

[0028] Figure 3 This is an example diagram of a landslide identification device for canyon-type reservoir areas provided according to an embodiment of this application;

[0029] Figure 4 This is a schematic diagram of the structure of an electronic device provided according to an embodiment of this application. Detailed Implementation

[0030] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0031] Landslide susceptibility refers to a quantitative or qualitative assessment of the classification, volume (or area), and spatial distribution of landslides that exist or may occur in a certain area.

[0032] Currently, landslide susceptibility mapping methods both domestically and internationally are mainly classified into three categories: empirical models, statistical analysis models, and machine learning models. Empirical models rely on expert knowledge and specific factors to evaluate landslide susceptibility within a region; examples include expert knowledge-based methods and the analytic hierarchy process (AHP). Statistical models quantitatively analyze the relationship between landslides and influencing factors; commonly used models include information value models, frequency ratio models, and weighted evidence methods. The relationship between landslides and influencing factors is complex, and machine learning models can effectively explain the nonlinear relationships involved, leading to their widespread application in landslide susceptibility assessment. Among machine learning models, random forests and gradient boosting machines often yield more effective landslide susceptibility assessment results. However, due to the complex mechanisms of landslide occurrence, a single model often struggles to represent the landslide susceptibility status of a region. Previous landslide susceptibility studies often relied on historical landslide lists as a reference, which may have affected the assessment results due to incomplete landslide data.

[0033] To this end, this application provides a method for identifying landslides in canyon-type reservoir areas. In this method, target deformation points in the canyon-type reservoir area can be determined based on radar data and topography, and the deformation rate of the target landslide deformation points can be calculated. Then, based on the deformation rate, a spatial clustering relationship of the target landslide deformation points is generated. Based on the deformation rate and spatial clustering relationship, landslide areas in the canyon-type reservoir area can be accurately identified, improving the accuracy of landslide area identification. Furthermore, the susceptibility level of the landslide area can be assessed, providing a basis for subsequent landslide disaster early warning.

[0034] Specifically, Figure 1 This is a flowchart illustrating a landslide identification method for a canyon-type reservoir area provided in an embodiment of this application.

[0035] like Figure 1 As shown, the method for identifying landslides in the canyon-type reservoir area includes the following steps:

[0036] In step S101, radar data and topographic features of the canyon-type reservoir area are acquired.

[0037] Among them, radar data can be obtained through satellite or ground radar systems, such as long-term time-series data of C-band synthetic aperture radar SAR using Sentinel-1; topography can be obtained through geographic information system GIS, and topography includes, for example, hills, mountains, and forests.

[0038] It is understood that the embodiments of this application can acquire radar data and topography of canyon-type reservoir areas for subsequent landslide identification.

[0039] In step S102, the target landslide deformation point in the canyon-type reservoir area is determined based on radar data and topography, and the deformation rate of the target landslide deformation point is calculated.

[0040] Among them, the target landslide deformation point is the deformation point with obvious deformation characteristics.

[0041] It is understood that, according to the embodiments of this application, the target landslide deformation point in the canyon-type reservoir area can be determined based on radar data and topographic data, and the deformation rate of the target landslide deformation point can be calculated in order to identify the landslide area in the future.

[0042] In this embodiment of the application, determining the target landslide deformation point in the canyon-type reservoir area based on radar data and topography includes: processing radar data to obtain multiple landslide deformation points and corresponding deformation rates in the canyon-type reservoir area; and filtering multiple landslide deformation points based on topography and deformation rates to obtain the target landslide deformation point in the canyon-type reservoir area.

[0043] Among them, SBAS-InSAR can be used to process radar data.

[0044] It is understood that, according to the embodiments of this application, multiple landslide deformation points and corresponding deformation rates in the canyon-type reservoir area can be obtained by processing radar data, and then multiple landslide deformation points can be screened based on topography and deformation rate, unreliable deformation points can be removed, and the target landslide deformation points in the canyon-type reservoir area can be obtained.

[0045] For example, when performing SBAS-InSAR processing on radar data, a digital elevation model is used to eliminate the influence of terrain phase, and POD precise orbit data is used for orbit correction to improve the accuracy of deformation monitoring. The obtained deformation results are converted into vector point format, and deformation points distributed in forests and rivers are removed to eliminate the error caused by their low accuracy, thus obtaining the target deformation points. Furthermore, during the processing of radar data, the RI index can be used for visualization analysis of radar data. For areas where geometric distortion is caused by the visibility of SAR data, the RI index is used for quantitative assessment to remove deformation points in radar image perspective contraction, overlay, and shadow areas, thereby enhancing the reliability of deformation point data.

[0046] In this embodiment of the application, the deformation rate of the target landslide deformation point is calculated, including obtaining the deformation rate in the line of sight corresponding to the target deformation point; and converting the deformation rate in the line of sight into a deformation rate perpendicular to the ground based on the radar geometry and topography.

[0047] Among them, the deformation rate in the line-of-sight direction is the deformation rate in the radar direction.

[0048] It is understood that the embodiments of this application can obtain the deformation rate of the line of sight corresponding to the target deformation point through processed radar data, and convert the deformation rate of the line of sight of the deformation vector point into a deformation rate perpendicular to the ground based on the radar geometry and topography, so as to more accurately describe the real deformation of the landslide deformation point and improve the reliability of the landslide deformation point.

[0049] In step S103, the spatial clustering relationship of the target landslide deformation points is generated based on the deformation rate, and the landslide area in the canyon-type reservoir area is identified based on the deformation rate and the spatial clustering relationship.

[0050] It is understood that the embodiments of this application can generate the spatial clustering relationship of the target landslide deformation points based on the deformation rate, and identify landslide areas in canyon-type reservoir areas according to the deformation rate and spatial clustering relationship. For example, areas with larger deformation rates and higher spatial clustering are identified as landslide areas. The spatial clustering relationship can reveal the spatial distribution pattern of landslide activity, and the identification of landslide areas is not limited by historical landslide inventory data.

[0051] In this embodiment of the application, generating the spatial clustering relationship of the deformation points of the target landslide based on the deformation rate includes: calculating the spatial distribution pattern of the deformation points of the target landslide based on the deformation rate; and generating the spatial clustering relationship of the deformation points of the target landslide based on the spatial distribution pattern.

[0052] It is understood that the embodiments of this application can calculate the spatial distribution pattern of the deformation points of the target landslide based on the deformation rate, and then generate the spatial clustering relationship of the deformation points of the target landslide based on the spatial distribution pattern, as follows:

[0053] Generate a spatial distribution map of deformation rates to visualize the distribution of deformation points; use spatial clustering algorithms (such as DBSCAN, K-means, etc.) to perform cluster analysis on the deformation rate data to identify groups of deformation points with similar deformation rates and spatial proximity; calculate the density of deformation points around each deformation point to reflect the degree of spatial aggregation of deformation points.

[0054] In this embodiment of the application, after identifying landslide areas in the canyon-type reservoir area based on deformation rate and spatial clustering relationship, the method further includes: assessing the landslide susceptibility level of the landslide area; and issuing early warnings or reminders based on the landslide susceptibility level.

[0055] The landslide susceptibility level can be preset according to specific circumstances, such as being set into five categories: "very low", "low", "medium", "high" and "very high".

[0056] It is understood that, after identifying a landslide area, the embodiments of this application can assess its landslide susceptibility level and issue early warnings or reminders based on the landslide susceptibility level, which can provide an important basis for disaster management and help decision-makers take appropriate preventive measures.

[0057] In this embodiment of the application, assessing the landslide susceptibility level of a landslide area includes: obtaining at least one landslide influencing factor of the landslide area; calculating the normalized information value and weight of each landslide influencing factor; and assessing the landslide susceptibility level of the landslide area based on the normalized information value and weight.

[0058] Landslide influencing factors may include elevation, slope, aspect, surface relief, NDVI, NDWI, water body distance, and fault distance.

[0059] It is understood that the embodiments of this application can calculate the normalized information value and weight of each landslide influencing factor in the landslide area, and assess the landslide susceptibility level of the landslide area based on the normalized information value and weight, comprehensively considering the influence of multiple landslide influencing factors on landslide susceptibility, so as to achieve a more comprehensive assessment of the landslide susceptibility level.

[0060] In this embodiment of the application, the normalized information value and weight of each landslide impact factor are calculated, including: classifying each landslide impact factor into levels and using an information value model to calculate the information value of each level of each landslide impact factor; unifying the information values ​​of each level of each landslide impact factor to obtain a normalized information value; using the classification characteristics of the landslide area and each landslide impact factor as samples; and performing a binary classification of the samples as landslide and non-landslide to obtain the weight of each landslide impact factor.

[0061] Normalization can be performed based on the arctangent function; binary classification can be achieved using methods such as random forest and extreme gradient boosting.

[0062] It is understood that the embodiments of this application can classify each landslide impact factor into levels, calculate the information value of each level using the information value model, unify the information value of each level to obtain a normalized information value, use the classification characteristics of landslide areas and each landslide impact factor as samples, perform landslide and non-landslide binary classification on the samples to obtain the weight of each landslide impact factor, thereby improving the scientificity and accuracy of the assessment.

[0063] Specifically, the landslide identification method for canyon-type reservoir areas in this application mainly consists of two parts: landslide area identification and landslide susceptibility assessment. Reliable landslide deformation points are selected based on radar data and topographic features, and their effective landslide deformation rates are calculated. The spatial clustering relationship between deformation points and their deformation rates is used to identify landslide areas. Using the identified landslide areas as a list, a new landslide susceptibility assessment model is proposed. This model uses the arctangent function to improve the information value model and couples it with random forest and extreme gradient boosting methods respectively. This completes the landslide susceptibility mapping and assessment of the target reservoir area, revealing the characteristics of landslide occurrence in the reservoir area and the contribution of different geographical factors, providing a reference for landslide risk early warning and management in the reservoir area. Figure 2 As shown, it includes the following steps:

[0064] Step S1: Reliable Deformation Point Acquisition in the Target Area. To obtain reliable landslide deformation points, SBAS-InSAR processing is performed using long-term time-series C-band synthetic aperture radar (SAR) data from Sentinel-1. Deformation points are then filtered based on factors such as terrain and land cover. The effective landslide deformation rate is calculated to further obtain reliable landslide deformation rate points. This part includes three components: SBAS-InSAR processing, visibility analysis based on the RI index, and landslide point filtering.

[0065] Step S1.1: SBAS-InSAR Processing and Result Optimization. Based on landslide monitoring requirements, Sentinel-1 imagery data within the target area's monitoring timeframe is used for SBAS-InSAR processing. During processing, a digital elevation model (DEM) is used to eliminate the influence of terrain phase, and POD precise orbit data is used for orbit correction to improve the accuracy of deformation monitoring. The obtained deformation results are converted to vector point format, and deformation points distributed in forest areas and rivers are removed to eliminate errors caused by their low accuracy.

[0066] Step S1.2: Visibility analysis based on the RI index. For areas whose geometric distortion is caused by the visibility of SAR data, the RI index is used for quantitative assessment to remove distortion points in radar image perspective shrinkage, overlay, and shadow areas, thereby enhancing the reliability of distortion point data.

[0067] Step S1.3: Landslide Point Screening. Having obtained the deformation points and deformation rates of the target area through the above steps, to obtain landslide deformation points and more accurately describe the true deformation of the landslide, the effective landslide deformation rate is recalculated based on the deformation rate along the line of sight obtained from the deformation vector points. On this basis, reliable landslide deformation points in the target area are obtained through screening based on the negative value of the landslide deformation and the stability threshold.

[0068] Step S2: Landslide Area Identification by Integrating Landslide Point Attribute Data and Spatial Clustering Relationships. The effective landslide deformation rate and its spatial clustering relationship of the obtained reliable landslide deformation points can reflect the stability of the regional landslides. Landslide areas are identified by filtering and clustering the landslide deformation points based on spatial relationships. First, the spatial distribution pattern of landslide points is calculated based on the effective landslide deformation rate attribute data to obtain their spatial clustering and dependencies. For landslides to occur, areas with higher deformation rates and denser clustering are more prone to landslides. Therefore, randomly distributed deformation points are removed based on the results. The distribution density of the removed landslide vector points in the entire reservoir study area is calculated to more intuitively show their spatial clustering degree. The clustering results are divided into two categories using the natural discontinuity method. The category with higher values ​​indicates that the overall landslide deformation is more severe and is considered as the identified landslide area.

[0069] Step S3: Landslide Susceptibility Mapping. A new landslide susceptibility assessment model is proposed, which improves the information value model using the arctangent function. Then, random forest and extreme gradient boosting methods are used to weight the improved information value model to calculate the landslide susceptibility of the target reservoir area. This includes the following three steps.

[0070] Step S3.1: Calculate the normalized information content (NIP) of different subcategories of landslide influencing factors. Collect multiple geographic factor data for the study area and reclassify them in ArcGIS using methods including natural breakpoint method and artificial threshold method. Based on the identified landslide areas, use the information value model to calculate the NIP of each level of different geographic factors. This NIP can measure the degree of their impact on landslide occurrence. When no landslide occurs in a certain level of influencing factor, the calculated NIP is negative infinity, making subsequent landslide susceptibility calculations impossible. To improve data stability, the NIP model is improved by using the arctangent function to normalize the NIP, thereby calculating the normalized NIP of different subcategories of each geographic factor.

[0071] Step S3.2: Calculate the weight of each landslide influencing factor on landslide occurrence. Using the landslide areas identified in Step S2 and their corresponding hierarchical characteristics of multiple geographical factors as samples, perform "landslide-non-landslide" binary classification using random forest and extreme gradient boosting methods respectively, and verify the accuracy of the classification. Obtain the importance of different landslide influencing factors, and analyze the landslide occurrence in the target reservoir area according to the importance.

[0072] Step S3.3: Calculate landslide susceptibility based on weighted normalized information content. The weighted normalized information content is calculated using the normalized information content obtained in step S3.1 and the corresponding importance obtained in step S3.2, and is used to characterize landslide susceptibility. The natural discontinuity method is used to classify the obtained landslide susceptibility into five categories, represented by "very low," "low," "medium," "high," and "very high," respectively, thus obtaining the landslide susceptibility distribution of the target area, facilitating early warning and management.

[0073] The following is a specific embodiment describing the landslide identification method for canyon-type reservoir areas of this application. The target areas are selected as Hekou Village Reservoir located at the border of Shanxi and Henan provinces and Baihetan Reservoir in Sichuan Province. The image data selected are Sentinel-1, Sentinel-2, and SRTM DEM, among other data. The programming tool selected is Python. The specific steps include:

[0074] Step S01: Obtain reliable landslide deformation points in the Hekou Village and Baihetan Reservoir areas. Download Sentinel-1 image data for the Hekou Village Reservoir area from January 2020 to December 2020 and the Baihetan Reservoir area from July 2021 to April 2023, and obtain reliable landslide deformation points within their respective areas according to the following.

[0075] Step S011: SBAS-InSAR Processing. Based on landslide monitoring requirements, SBAS-InSAR processing was performed on the Sentinel-1 image data downloaded from the Hekou Village and Baihetan Reservoir areas. Simultaneously, SRTM DEM data with a resolution of 30m for the corresponding areas was used to eliminate the influence of topographic phase, and POD precise orbit data was used for orbit correction. To facilitate subsequent spatial analysis, the deformation results were converted into vector point format. Since the accuracy of the obtained deformation points is affected by high vegetation density and cover, the European Space Agency's land cover product was used to remove deformation points distributed in woodlands and rivers.

[0076] Step S012: Visibility Analysis Based on the RI Index. In SAR system imaging, elevation changes and the angular relationship between antenna direction and surface morphology can cause geometric distortions such as perspective contraction, overlay, and shadows in SAR images. Areas with good visibility are identified using satellite azimuth, incident angle, and slope and aspect data obtained from DEM data. Deformation points in the more visible areas of Hekou Village and Baihetan Reservoir area are preserved to enhance the reliability of deformation point data.

[0077] Step S013: Landslide Point Screening. Following the above steps, deformation points with good LOS visibility and their deformation rates were obtained in Hekou Village and the Baihetan Reservoir area. To more accurately describe the actual landslide deformation, the effective landslide deformation rate was recalculated based on the obtained LOS deformation rate. According to the characteristics of landslide occurrence, deformation points with negative effective landslide deformation rates were retained in both target areas, and their standard deviation was calculated as a stability threshold. Deformation points with deformation rates exceeding the stability threshold were considered reliable landslide deformation points identified in Hekou Village and the Baihetan Reservoir area.

[0078] Step S02: Landslide Area Identification by Integrating Attribute Data and Spatial Clustering Relationships of Landslide Points. The effective landslide deformation rate characteristics and their spatial clustering relationships among reliable landslide deformation points obtained after screening can reflect the stability of landslides in the reservoir area. Therefore, landslide areas are identified by filtering and clustering landslide deformation points based on spatial relationships. First, the spatial distribution pattern of landslide points is calculated based on the attribute data of effective landslide deformation rate, obtaining their spatial clustering and dependencies. For landslides to occur, areas with higher deformation rates and denser clustering are more prone to landslides. Therefore, spatial filtering is performed based on the analysis results to remove randomly distributed deformation points. The distribution density of the removed landslide vector points throughout the study area is calculated to more intuitively display their spatial clustering degree. The clustering results are divided into two categories using the natural discontinuity method. The category with higher values ​​indicates that the overall landslide deformation is more severe and is considered the identified landslide area.

[0079] Step S03: Landslide Susceptibility Mapping. A new landslide susceptibility assessment model is proposed, which improves the information value model using the arctangent function and couples it with random forest and the ultimate gradient boosting method to calculate the landslide susceptibility of the target reservoir area. Based on the characteristics of the reservoir area, elevation, slope, aspect, surface relief, NDVI, NDWI, water body distance, and fault distance are selected as geographical factors affecting landslides and reclassified. The improved information value model is used to calculate the normalized information content of landslide occurrence. Random forest and the ultimate gradient boosting method are used to weight the improved information value model to calculate the landslide susceptibility of the Hekou Village and Baihetan Reservoir areas. This includes the following three steps:

[0080] Step S031: Calculate the normalized information content of different subcategories of landslide influencing factors. The landslide influencing factors in Hekou Village and Baihetan were classified separately using methods including the natural discontinuity method and artificial threshold division. Information values ​​for each level of different landslide influencing factors were then calculated to measure their impact on landslide occurrence. Because the information content of the classified influencing factors in the Hekou Village reservoir area included subcategories where no landslides had occurred, resulting in negative infinity for elevation, NDVI, and fault distance, it was impossible to make subsequent landslide susceptibility assessments. To improve data stability, the arctangent function was used to normalize the information content.

[0081] Step S032: Calculate the importance of each landslide influencing factor to landslide occurrence. Using the landslide areas of Hekou Village and Baihetan identified in Step S02 and their corresponding landslide influencing factors as samples, random forest and extreme gradient boosting methods are used to perform binary classification of landslides and non-landslides to obtain the importance weights of different landslide influencing factors. ROC curves, AUC, Precision, Recall, and F1 indices are used to verify the accuracy of the classification. The landslide occurrence in the Hekou Village and Baihetan reservoir areas is evaluated based on the importance obtained from each influencing factor. In the canyon-type landslide reservoir areas such as Hekou Village and Baihetan, slope aspect has a significant impact on landslide occurrence and is a key factor that needs to be focused on.

[0082] Step S033: Calculate landslide susceptibility based on normalized information content and importance. The weighted normalized information content for Hekou Village and Baihetan Reservoir area is calculated using the normalized information content obtained in Step S031 and the importance of the corresponding influencing factors obtained in Step S032. This weighted normalized information content characterizes landslide susceptibility; a higher value indicates greater susceptibility. The natural discontinuity method is used to classify the obtained landslide susceptibility into five categories, represented by "very low," "low," "medium," "high," and "very high." Visualization and analysis reveal that areas with high landslide susceptibility are mainly concentrated near water bodies, and the susceptibility gradually decreases with increasing distance from the water body. This provides a reference for landslide occurrence in canyon-type reservoir areas such as Hekou Village and Baihetan, facilitating timely early warning and management.

[0083] In summary, the landslide identification method for canyon-type reservoir areas in this application provides an integrated method for identification and assessment, which can accurately perform landslide early warning and comprehensive assessment. It introduces deformation points and their attribute data and spatial clustering relationships to identify landslide areas, and is not limited by historical landslide inventory data when assessing landslide susceptibility. Based on multiple geographical factors, the information value model is improved using the arctangent function, and then coupled with random forest and extreme gradient boosting methods to evaluate and compare the landslide susceptibility of reservoir areas.

[0084] The landslide identification method for canyon-type reservoir areas proposed in this application can determine the target deformation points of the canyon-type reservoir area based on radar data and topography, calculate the deformation rate of the target landslide deformation points, and then generate the spatial clustering relationship of the target landslide deformation points based on the deformation rate. Based on the deformation rate and spatial clustering relationship, the landslide area in the canyon-type reservoir area can be accurately identified, improving the accuracy of landslide area identification and providing a basis for subsequent landslide disaster early warning.

[0085] Next, referring to the accompanying drawings, we describe the landslide identification device for canyon-type reservoir areas proposed according to the embodiments of this application.

[0086] Figure 3 This is a block diagram of a landslide identification device for a canyon-type reservoir area according to an embodiment of this application.

[0087] like Figure 3 As shown, the landslide identification device 10 for the canyon-type reservoir area includes: an acquisition module 100, a determination module 200, and an identification module 300.

[0088] The acquisition module 100 is used to acquire radar data and topography of the canyon-type reservoir area; the determination module 200 is used to determine the target landslide deformation points in the canyon-type reservoir area based on the radar data and topography, and calculate the deformation rate of the target landslide deformation points; the identification module 300 is used to generate the spatial clustering relationship of the target landslide deformation points based on the deformation rate, and identify the landslide area in the canyon-type reservoir area based on the deformation rate and the spatial clustering relationship.

[0089] In this embodiment of the application, the apparatus 10 further includes an evaluation module.

[0090] The assessment module is used to assess the landslide susceptibility level of landslide areas after identifying landslide areas in canyon-type reservoir areas based on deformation rate and spatial clustering relationship; and to issue early warnings or alerts based on the landslide susceptibility level.

[0091] In this embodiment of the application, the determining module 200 is further configured to: obtain the deformation rate in the line of sight corresponding to the target deformation point; and convert the deformation rate in the line of sight into a deformation rate perpendicular to the ground based on the radar geometry and terrain.

[0092] In this embodiment of the application, the determining module 200 is further used to: process radar data to obtain multiple landslide deformation points and corresponding deformation rates in the canyon-type reservoir area; and filter multiple landslide deformation points based on topography and deformation rate to obtain the target landslide deformation point in the canyon-type reservoir area.

[0093] In this embodiment of the application, the identification module 100 is further used to: calculate the spatial distribution pattern of the deformation points of the target landslide based on the deformation rate; and generate the spatial clustering relationship of the deformation points of the target landslide based on the spatial distribution pattern.

[0094] In this embodiment of the application, the evaluation module is further used to: classify each landslide impact factor into levels, and calculate the information value of each level of each landslide impact factor using the information value model; unify the information values ​​of each level of each landslide impact factor to obtain a normalized information value; use the classification characteristics of the landslide area and each landslide impact factor as samples; and perform landslide and non-landslide binary classification on the samples to obtain the weight of each landslide impact factor.

[0095] It should be noted that the foregoing explanation of the embodiment of the landslide identification method for canyon-type reservoir areas also applies to the landslide identification device for canyon-type reservoir areas in this embodiment, and will not be repeated here.

[0096] The landslide identification device for canyon-type reservoir areas proposed in this application can determine the target deformation points of the canyon-type reservoir area based on radar data and topography, calculate the deformation rate of the target landslide deformation points, and then generate the spatial clustering relationship of the target landslide deformation points based on the deformation rate. Based on the deformation rate and spatial clustering relationship, the device can accurately identify landslide areas in the canyon-type reservoir area, improve the accuracy of landslide area identification, and provide a basis for subsequent landslide disaster early warning.

[0097] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may include:

[0098] The memory 401, the processor 402, and the computer program stored on the memory 401 and capable of running on the processor 402.

[0099] When the processor 402 executes the program, it implements the landslide identification method for canyon-type reservoir areas provided in the above embodiments.

[0100] Furthermore, the identification also includes:

[0101] Communication interface 403 is used for communication between memory 401 and processor 402.

[0102] The memory 401 is used to store computer programs that can run on the processor 402.

[0103] The memory 401 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0104] If the memory 401, processor 402, and communication interface 403 are implemented independently, then the communication interface 403, memory 401, and processor 402 can be interconnected via a bus to complete communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be divided into address buses, data buses, control buses, etc. For ease of representation, Figure 4 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0105] Optionally, in a specific implementation, if the memory 401, processor 402, and communication interface 403 are integrated on a single chip, then the memory 401, processor 402, and communication interface 403 can communicate with each other through an internal interface.

[0106] Processor 402 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application.

[0107] This application also provides a computer-readable storage medium storing a computer program or instructions thereon, which, when executed by a processor, implements the above-described method for identifying landslides in canyon-type reservoir areas.

[0108] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0109] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0110] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or N executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0111] It should be understood that the various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or more of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0112] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

Claims

1. A method for identifying landslides in canyon-type reservoir areas, characterized in that, Includes the following steps: Acquire temporal radar interferometry data and topographic features of the canyon-type reservoir area; The target landslide deformation point in the canyon-type reservoir area is determined based on the time-series radar interferometry data and the topography, and the deformation rate of the target landslide deformation point is calculated. Based on the deformation rate, a spatial clustering relationship of the target landslide deformation points is generated, and landslide areas in the canyon-type reservoir area are identified based on the deformation rate and the spatial clustering relationship. After identifying landslide areas in the canyon-type reservoir area based on the deformation rate and the spatial clustering relationship, the process includes: evaluating the susceptibility level of the landslide areas using a landslide susceptibility assessment model. The landslide susceptibility assessment model processing includes: obtaining at least one landslide influencing factor for the landslide area; classifying each landslide influencing factor into levels and calculating the information value of each level of each landslide influencing factor using an information value model; unifying the information values ​​of each level of each landslide influencing factor using an arctangent function to obtain a normalized information value; using the classification characteristics of the landslide area and each landslide influencing factor as samples; performing landslide and non-landslide binary classification on the samples using a random forest method and a limit gradient boosting method to obtain the weight of each landslide influencing factor; and evaluating the landslide susceptibility level of the landslide area based on the normalized information value and the weight.

2. The method for identifying landslides in canyon-type reservoir areas according to claim 1, characterized in that, After evaluating the susceptibility level of the landslide area using a landslide susceptibility assessment model, the following steps are also included: Early warnings or alerts are issued based on the landslide susceptibility level.

3. The method for identifying landslides in canyon-type reservoir areas according to claim 1, characterized in that, The calculation of the deformation rate of the target landslide deformation point includes: Obtain the deformation rate along the radar line of sight corresponding to the target deformation point; Based on the radar's geometry and the terrain, the deformation rate along the line of sight is converted into a deformation rate perpendicular to the ground.

4. The method for identifying landslides in canyon-type reservoir areas according to claim 2, characterized in that, The step of determining the target landslide deformation point in the canyon-type reservoir area based on the time-series radar interferometry data and the topography includes: The radar data is processed to determine multiple landslide deformation points and their corresponding deformation rates in the canyon-type reservoir area. Based on the topography and deformation rate, the multiple landslide deformation points are screened to obtain the target landslide deformation points in the canyon-type reservoir area.

5. The method for identifying landslides in canyon-type reservoir areas according to claim 1, characterized in that, The generation of the spatial clustering relationship of the target landslide deformation points based on the deformation rate includes: Calculate the spatial distribution pattern of the deformation points of the target landslide based on the deformation rate; The spatial clustering relationship of the target landslide deformation points is generated based on the spatial distribution pattern.

6. A landslide identification device for canyon-type reservoir areas, characterized in that, include: The acquisition module is used to acquire time-series radar interferometry data and topographic features of the canyon-type reservoir area; The determination module is used to determine the target landslide deformation point in the canyon-type reservoir area based on the time-series radar interferometry data and the topography, and to calculate the deformation rate of the target landslide deformation point; The identification module is used to generate the spatial clustering relationship of the deformation points of the target landslide based on the deformation rate, and to identify the landslide area in the canyon-type reservoir area based on the deformation rate and the spatial clustering relationship. After identifying landslide areas in the canyon-type reservoir area based on the deformation rate and the spatial clustering relationship, the process includes: evaluating the susceptibility level of the landslide areas using a landslide susceptibility assessment model. The landslide susceptibility assessment model processing includes: obtaining at least one landslide influencing factor for the landslide area; classifying each landslide influencing factor into levels and calculating the information value of each level of each landslide influencing factor using an information value model; unifying the information values ​​of each level of each landslide influencing factor using an arctangent function to obtain a normalized information value; using the classification characteristics of the landslide area and each landslide influencing factor as samples; performing landslide and non-landslide binary classification on the samples using a random forest method and a limit gradient boosting method to obtain the weight of each landslide influencing factor; and evaluating the landslide susceptibility level of the landslide area based on the normalized information value and the weight.

7. An electronic device, characterized in that, include: The method includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the landslide identification method for canyon-type reservoir areas as described in any one of claims 1-5.

8. A computer-readable storage medium having a computer program or instructions stored thereon, characterized in that, The computer program or instructions are executed by a processor to implement the landslide identification method for canyon-type reservoir areas as described in any one of claims 1-5.

Images