A method and device for constructing a loess landslide crack evolution trend map

Abstract

The present invention discloses a method and device for constructing a loess landslide fissure evolution trend map, relating to the technical field of early-stage micro-topography identification of landslide hazards. The method comprises: acquiring multi-source remote sensing image data of a loess landslide; acquiring reconstruction data of the loess landslide based on the multi-source remote sensing image data, and extracting fissure features of the loess landslide based on the reconstruction data; establishing a feature extraction atlas library based on the multi-source remote sensing image data, and obtaining a classification result of the fissure features based on the feature extraction atlas library, the fissure features, and the location parameters of the fissure features; identifying the fissure features collected at multiple time periods based on the classification results of the fissure features, obtaining identification results of the fissure features at multiple time periods; and constructing a fissure evolution trend map of the loess landslide based on the identification results of the fissure features at multiple time periods. Through the above scheme, the location of potential loess landslide hazards can be predicted.

Description

Technical Field

[0001] The present invention relates to the technical field of early-stage micro-topography identification of landslide disasters, and in particular to a method and device for constructing a loess landslide fissure evolution trend map. Background Art

[0002] Landslides are one of the major natural disasters worldwide, causing significant economic losses and casualties, and severely endangering people's lives and property. Landslides require certain geological conditions to occur. Most loess landslides exhibit specific precursory features (such as cracks, steep slopes, and sinkholes) before they occur. These slopes, displaying these precursory features, are often significant landslide hazard risk points.

[0003] Existing methods for detecting landslide precursors mainly involve using drones equipped with optical sensors to obtain real texture information of the landslide area and identify and catalog surface cracks; or using Bayesian probability calculation formulas to predict the probability of another landslide after multiple landslides have occurred; or using artificial neural networks to predict the location and time of landslide occurrence.

[0004] However, existing technologies are unable to evaluate the evolutionary relationship between multiple landslide disasters, resulting in low accuracy in predicting the location of landslide disasters. Summary of the Invention

[0005] Based on this, it is necessary to provide a method and device for constructing a loess landslide crack evolution trend map to address the problems of low detection accuracy and slow detection speed in existing surface defect detection methods for polysilicon production equipment.

[0006] The present invention adopts the following technical solutions:

[0007] In a first aspect, the present invention provides a method for constructing a loess landslide fissure evolution trend map, the method comprising:

[0008] Acquiring multi-source remote sensing image data of a loess landslide, wherein the multi-source remote sensing image data includes optical image data, laser point cloud data, and surface temperature distribution data;

[0009] Acquiring reconstruction data of the loess landslide based on the multi-source remote sensing image data, and extracting crack characteristics of the loess landslide based on the reconstruction data;

[0010] Establishing a feature extraction atlas library based on the multi-source remote sensing image data, and obtaining a classification result of the crack feature based on the feature extraction atlas library, the crack feature, and the position parameters of the crack feature;

[0011] Identifying the crack features collected at multiple periods according to the classification results of the crack features to obtain identification results of the crack features at the multiple periods;

[0012] A crack evolution trend map of the loess landslide is constructed based on the identification results of the crack characteristics of the multiple periods.

[0013] Furthermore, obtaining the reconstructed data of the loess landslide based on the multi-source remote sensing image data specifically includes:

[0014] Acquiring texture information of a fissure region of the loess landslide based on the optical image data, and reconstructing an optical image of the loess landslide according to the texture information;

[0015] generating a three-dimensional reconstruction model of the ground and exposed areas of the loess landslide based on the laser point cloud data;

[0016] The thermal response characteristics of the fissure area are reconstructed based on the surface temperature distribution data, and a thermal infrared model of the loess landslide is obtained according to the thermal response characteristics.

[0017] Furthermore, extracting the fissure characteristics of the loess landslide based on the reconstructed data specifically includes:

[0018] establishing an image dataset about the fissure based on the reconstructed data;

[0019] Performing grayscale processing on the image dataset using an edge detection algorithm, and obtaining an edge dataset of the crack based on the grayscale processed image dataset;

[0020] A segmentation algorithm is used to extract features from the edge data set of the crack to obtain the crack features of the loess landslide.

[0021] Furthermore, a feature extraction atlas library is established based on the multi-source remote sensing image data, and a classification result of the crack feature is obtained based on the feature extraction atlas library, the crack feature, and the position parameter of the crack feature, specifically including:

[0022] Building a fracture sample library based on the multi-source remote sensing image data;

[0023] Extracting characteristic parameters of each crack in the crack sample library, wherein the characteristic parameters include crack position, crack morphology and temperature;

[0024] The fractures in the fracture sample library are classified according to their characteristic parameters, and the fractures in the classified fracture sample library are combined with the fracture characteristics of the loess landslide and the position parameters of the loess landslide to obtain a classification result of the fracture characteristics of the loess landslide.

[0025] Furthermore, after obtaining the classification result of the fissure characteristics of the loess landslide, the method further includes:

[0026] Obtaining physical parameters of each of the cracks, converting the physical parameters of each of the cracks into standardized values, and obtaining characteristic vectors corresponding to the physical parameters of each of the cracks based on the standardized values, wherein the physical parameters include length, width, depth, direction, roughness, slope, and temperature;

[0027] The characteristic vectors corresponding to the physical parameters of each crack are standardized and stored using a database system to obtain a standardized crack sample library, and the crack characteristics of the loess landslide are re-identified using the standardized crack sample library to obtain an optimized crack characteristic classification result.

[0028] Furthermore, based on the classification results of the crack features, the crack features collected at multiple periods are identified to obtain identification results of the crack features at the multiple periods, specifically including:

[0029] respectively calculating the similarity between each of the crack features in the plurality of periods and each type of crack feature in the classification results of the crack features;

[0030] The respective crack features are identified according to the similarity to obtain identification results of the crack features of the multiple periods.

[0031] Furthermore, a fissure evolution trend map of the loess landslide is constructed based on the identification results of the fissure characteristics in the multiple periods, specifically including:

[0032] Converting the data corresponding to the identification results of the crack features of the multiple periods into a spatial data format to obtain spatial data corresponding to the crack features of the multiple periods, and marking the spatial data corresponding to the crack features of the multiple periods with time stamp information;

[0033] Loading the fracture features of multiple periods marked with time stamp information into a geographic information system for spatial registration, and setting different colors and transparencies for the fracture features of multiple periods marked with time stamp information;

[0034] The fissure characteristics of the multiple periods are superimposed on the map layout of the geographic information system, and map elements are added to the map layout to obtain the loess landslide fissure evolution trend map.

[0035] In a second aspect, the present invention provides a device for constructing a loess landslide fissure evolution trend map, comprising:

[0036] A first acquisition module is used to acquire multi-source remote sensing image data of loess landslides, wherein the multi-source remote sensing image data includes optical image data, laser point cloud data and surface temperature distribution data;

[0037] a reconstruction module, configured to obtain reconstruction data of the loess landslide based on the multi-source remote sensing image data, and extract crack features of the loess landslide based on the reconstruction data;

[0038] A classification module is used to establish a feature extraction atlas library based on the multi-source remote sensing image data, and obtain a classification result of the crack feature based on the feature extraction atlas library, the crack feature and the position parameter of the crack feature;

[0039] an identification module, configured to identify the crack features collected at multiple periods according to the classification results of the crack features, and obtain identification results of the crack features at the multiple periods;

[0040] The second acquisition module is used to construct a crack evolution trend map of the loess landslide according to the identification results of the crack characteristics of the multiple periods.

[0041] The present invention provides a computer-readable storage medium, wherein the storage medium stores a computer program. When the computer program is executed by a processor, the method for constructing a loess landslide fissure evolution trend map is implemented.

[0042] The present invention provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and runnable on the processor. When the processor executes the program, a method for constructing a loess landslide fissure evolution trend map is implemented.

[0043] The at least one technical solution adopted by the present invention can achieve the following beneficial effects:

[0044] The present invention can obtain the crack characteristics of loess landslides based on multi-source remote sensing image data of loess landslides, establish a feature extraction atlas library for loess landslides based on the multi-source remote sensing image data of loess landslides, and obtain classification results of the crack characteristics based on the feature extraction atlas library, the crack characteristics, and the location parameters of the crack characteristics. Finally, based on the classification results of the crack characteristics, the crack characteristics collected at multiple time periods are identified to obtain identification results of the crack characteristics at multiple time periods. Based on the identification results of the crack characteristics at multiple time periods, a crack evolution trend map of the loess landslide is constructed. The crack maps of multiple time periods can be cross-compared to determine the crack development areas of loess landslides over a large area, preliminarily determine the location of loess landslide hidden danger areas, and accurately predict the occurrence location of loess landslides. Moreover, as the number of loess landslide crack characteristics in the crack evolution trend map increases, the prediction of the occurrence location of loess landslides is further improved. The crack evolution trend map can be used to evaluate the evolution relationship between multiple landslide disasters. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0046] Figure 1 A flow chart of a method for constructing a loess landslide fissure evolution trend map provided by the present invention;

[0047] Figure 2 A loess landslide fissure evolution trend map provided by the present invention;

[0048] Figure 3 A training curve diagram of a crack recognition model provided by the present invention;

[0049] Figure 4 A partial result diagram of crack identification provided by the present invention;

[0050] Figure 5 This is a general framework diagram of a method for constructing a loess landslide fissure evolution trend map provided by the present invention;

[0051] Figure 6 A schematic diagram of a device for constructing a loess landslide fissure evolution trend map provided by the present invention;

[0052] Figure 7 A schematic diagram of a computer device for implementing a method for constructing a loess landslide fissure evolution trend map provided by the present invention. DETAILED DESCRIPTION

[0053] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with specific embodiments of the present invention and corresponding drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0054] The server mentioned in the present invention can be a server installed on a business platform, or a device such as a desktop computer or laptop computer capable of executing the solution of the present invention. For ease of explanation, the following description will only use the server as the execution entity. The following, combined with the accompanying drawings, details the technical solutions provided by various embodiments of the present invention.

[0055] Figure 1 This is a flow chart of a method for constructing a loess landslide fissure evolution trend map in the present invention, which specifically includes the following steps:

[0056] S10: Acquire multi-source remote sensing image data of the loess landslide, where the multi-source remote sensing image data includes optical image data, laser point cloud data, and surface temperature distribution data.

[0057] In this embodiment, reference Figure 5 , multi-source remote sensing image data of loess landslides are obtained through multi-source sensors. Among them, the multi-source sensors are set on drones and include optical sensors, lidar sensors and thermal infrared sensors.

[0058] Alternatively, survey the terrain at the loess landslide site, select suitable takeoff and landing points for the drone, define a flight path, and set corresponding flight parameters. These parameters include altitude, speed, and direction, and set the optical sensor's camera mode to ensure clarity and integrity of the captured optical image data.

[0059] Among them, the route parameters of the lidar sensor are set to three-echo mode to ensure that the lidar has the effect of "penetrating" vegetation and measuring more surface cracks under vegetation cover; visible light coloring is turned on to ensure that the point cloud displays the actual terrain texture effect; only the lateral overlap rate needs to be set to obtain the complete mapping results of the laser point cloud data.

[0060] Optionally, after acquiring multi-source remote sensing image data, data alignment and registration, and data fusion of the multi-source remote sensing image data are also included. Data alignment and registration include:

[0061] Optical image data, laser point cloud data, and surface temperature distribution data are aligned to the same coordinate system using common control points to ensure their consistency in three-dimensional space. Data collected at different time points are time-synchronized to ensure consistency in the target areas of each data set, avoiding data offsets caused by time differences.

[0062] In this embodiment, the common control points of the data are key points used for accurate and effective registration between multiple images.

[0063] The fusion of multi-source remote sensing image data includes:

[0064] Fusion of laser point cloud data and optical image data: Laser point cloud data is used for 3D reconstruction to generate a Digital Elevation Model (DEM) or Digital Surface Model (DSM) of the landslide area. Optical image data is texture-mapped onto the laser point cloud data to generate a high-precision 3D reality model. Within the fused 3D reality model, geometric features of the crack area (such as location, length, width, and depth) are extracted through texture analysis and morphological methods (such as edge detection and segmentation algorithms).

[0065] Fusion of optical image data and surface temperature distribution data: Image processing methods such as histogram equalization and contrast adjustment are used to enhance crack features in the optical image data and surface temperature distribution data, especially temperature anomalies in the surface temperature distribution data. Image fusion techniques such as wavelet transform-based image fusion and principal component analysis (PCA) fusion are used to fuse the geometric information of the optical image data and surface temperature distribution data.

[0066] Fusion of laser point cloud data and surface temperature distribution data: A ground point classification algorithm is used to extract exposed ground points from the laser point cloud data and match them with temperature anomalies in the surface temperature distribution data. This temperature data extraction allows for further identification of cracks as they appear in thermal imaging.

[0067] Optionally, set the corresponding route parameters for the drone, including the following precautions:

[0068] The drone route should be set to fly steadily at the same altitude to avoid altitude fluctuations during flight, which may lead to unnecessary power consumption.

[0069] S20: Obtaining reconstruction data of the loess landslide based on the multi-source remote sensing image data, and extracting crack characteristics of the loess landslide based on the reconstruction data.

[0070] In this embodiment, the reconstruction data of the loess landslide is obtained based on multi-source remote sensing image data, specifically including:

[0071] The texture information of the crack area of ​​the loess landslide is obtained based on the optical image data, and the optical image of the loess landslide is reconstructed according to the texture information.

[0072] Generate a 3D reconstruction model of the ground and exposed areas of the loess landslide based on laser point cloud data.

[0073] The thermal response characteristics of the crack area are reconstructed based on the surface temperature distribution data, and the thermal infrared model of the loess landslide is obtained according to the thermal response characteristics.

[0074] In this embodiment, the actual position of each pixel point in the optical image data in three-dimensional space is inverted based on the position information and shooting parameters of the optical image data, and laser point cloud data of the fracture development area is generated. The landslide texture information in the optical image data is fused with the laser point cloud data, and the laser point cloud data is texture mapped to generate DOMs and a high-precision three-dimensional real-scene model.

[0075] Import the original laser point cloud data file in LDR format into software capable of laser point cloud reconstruction, set the point cloud density and effective distance, choose whether to enable point cloud accuracy optimization and point cloud smoothing, perform ground point classification, extract the elevation information of exposed ground points after removing noise points such as vegetation and buildings, and generate laser point cloud models and DEMs.

[0076] Import the surface temperature distribution data into software capable of infrared model reconstruction. Set the coordinate system and output file format. The software will automatically generate the corresponding thermal infrared model, which will visually display the temperature data within the landslide area.

[0077] Indicative, Figure 2 The present invention provides a loess landslide fissure evolution trend map, wherein: Figure 2 (a) is the optical three-dimensional model of the loess landslide. Figure 2 (b) is a 3D laser point cloud model. Figure 2 (c) in the figure is a three-dimensional thermal infrared model.

[0078] S30: establishing a feature extraction atlas library based on the multi-source remote sensing image data, and obtaining a classification result of the fracture feature based on the feature extraction atlas library, the fracture feature, and the position parameters of the fracture feature.

[0079] In this embodiment, the feature extraction atlas library refers to a collection of multiple images containing loess landslide fissure features.

[0080] S40: Identifying the collected crack features of multiple periods according to the classification results of the crack features to obtain identification results of the crack features of multiple periods.

[0081] In this embodiment, the classification results include crack features that belong to loess landslides and crack features that do not belong to loess landslides.

[0082] S50: Construct a crack evolution trend map of loess landslide based on the identification results of crack characteristics in multiple periods.

[0083] based on Figure 1 The method for constructing a loess landslide fissure evolution trend map is shown. It can obtain loess landslide fissure characteristics based on multi-source remote sensing image data of the loess landslide, establish a loess landslide feature extraction atlas library based on the multi-source remote sensing image data, and obtain fissure feature classification results based on the feature extraction atlas library, fissure characteristics, and fissure feature location parameters. Finally, based on the fissure feature classification results, the collected fissure characteristics from multiple time periods are identified to obtain identification results for the fissure characteristics of multiple time periods. Based on the identification results of the fissure characteristics of multiple time periods, a loess landslide fissure evolution trend map is constructed. The method can cross-compare the fissure maps from multiple time periods to determine the loess landslide fissure development zone over a large area, preliminarily determine the location of loess landslide potential risk areas, and accurately predict the occurrence location of loess landslides. Moreover, as the number of loess landslide fissure features in the fissure evolution trend map increases, the prediction of the location of loess landslides will be further improved, and the evolution relationship between multiple landslide disasters can be evaluated through the fissure evolution trend map.

[0084] When applying the method for constructing a loess landslide crack evolution trend map provided by the present invention, it is not necessary to Figure 1 The steps are executed in the order shown. The specific execution order of the steps can be determined according to needs, and the present invention does not limit this.

[0085] Furthermore, in one or more embodiments of the present invention, extracting crack features of a loess landslide based on the reconstructed data specifically includes:

[0086] An image dataset of cracks is established based on the reconstructed data.

[0087] The edge detection algorithm is used to perform grayscale processing on the image dataset, and the edge dataset of the crack is obtained based on the grayscale processed image dataset.

[0088] The segmentation algorithm is used to extract the features of the edge data set of the cracks and obtain the crack characteristics of the loess landslide.

[0089] In this embodiment, an image dataset of cracks is created from reconstructed data. This image dataset is then grayscale processed using an edge detection algorithm. Based on this grayscale processed image dataset, a crack edge dataset is obtained. Finally, a segmentation algorithm is used to extract features from this crack edge dataset, resulting in crack characteristics of the loess landslide. This improves the accuracy of crack feature extraction.

[0090] In addition, in one or more embodiments of the present invention, a feature extraction atlas library is established based on multi-source remote sensing image data, and a classification result of the fracture features is obtained based on the feature extraction atlas library, the fracture features, and the position parameters of the fracture features, specifically including:

[0091] Construct a fracture sample library based on multi-source remote sensing image data.

[0092] The characteristic parameters of each crack in the crack sample library are extracted, and the characteristic parameters include crack position, crack morphology and temperature.

[0093] Each crack in the crack sample library is classified according to its characteristic parameters, and the cracks in the classified crack sample library are combined with the crack characteristics and position parameters of the loess landslide to obtain the classification results of the crack characteristics of the loess landslide.

[0094] In this embodiment, classifying each crack according to its characteristic parameters means dividing each crack according to different crack positions, different crack shapes, and different temperature ranges, thereby obtaining cracks belonging to different crack positions, different crack shapes, and different temperature ranges.

[0095] It should be noted that when the characteristic parameters of one or more cracks in the crack sample library do not fall within the preset characteristic parameter range, the classification result of the one or more cracks is that the crack characteristics do not belong to loess landslide. The preset characteristic parameter range refers to the range of characteristic parameters of cracks belonging to loess landslide.

[0096] Specifically, different division intervals may be set for each characteristic parameter, so as to classify each crack according to the division interval in which the characteristic parameter of each crack lies.

[0097] In one or more embodiments of the present invention, after obtaining the classification results of the crack characteristics of the loess landslide, the method further includes:

[0098] The physical parameters of each crack are obtained and converted into standardized values. The characteristic vectors corresponding to the physical parameters of each crack are obtained based on the standardized values. The physical parameters include length, width, depth, direction, roughness, slope and temperature.

[0099] In this embodiment, the physical parameters of each fracture are obtained by extracting the fracture length, width, orientation, roughness, slope, and temperature based on the orthophoto image and 3D real-world model from the optical image data, and statistically classifying the fracture morphological characteristics. Combined with the 3D point cloud model and its derived DEMs, characteristic information such as fracture depth, roughness, dispersion, and slope is identified.

[0100] The temperature of the cracks is extracted by fusing optical image data with surface temperature distribution data, identifying the crack morphology based on the difference in thermal conductivity between the cracks and the surrounding loess, and calculating the crack temperature. This also describes the temperature differences between the cracks and surrounding homogeneous loess, landslide deposits, and other ground features.

[0101] The database system is used to standardize and store the characteristic vectors corresponding to the physical parameters of each crack to obtain a standardized crack sample library. The standardized crack sample library is then used to identify the crack characteristics of loess landslides again to obtain optimized crack feature classification results.

[0102] In this embodiment, a database system (e.g., a Structured Query Language (SQL) database) is used to store all fracture characteristic information in a standardized manner, forming a fracture sample library. Each fracture characteristic is stored as a field in a data table to ensure that the data can be easily queried and accessed.

[0103] This embodiment obtains a standardized crack sample library by standardizing and storing the characteristic vectors corresponding to the physical parameters of each crack, and uses the standardized crack sample library to re-identify the crack characteristics of loess landslides, which can further improve the recognition accuracy of the crack characteristics of loess landslides.

[0104] Furthermore, in one or more embodiments of the present invention, the fracture features collected at multiple periods are identified based on the fracture feature classification results, and the fracture feature identification results for the multiple periods are obtained, specifically including:

[0105] The similarity between each fracture feature in the fracture features of multiple periods and each type of fracture feature in the fracture feature classification results is calculated respectively.

[0106] In this embodiment, the similarity between each fracture feature in the fracture features of multiple periods and the fracture features and the existing fracture features in the fracture sample library is calculated to determine whether each fracture feature in the fracture features of multiple periods is already included in the fracture sample library; similarity judgment is performed on multiple parameters of the fracture features, and the judgment results are integrated to achieve efficient recognition.

[0107] Compare each fracture signature from multiple periods with the existing fracture signatures in the fracture sample library, including the following:

[0108] S1: Merge and deduplicate each fracture feature from multiple periods to remove redundant fracture information, ensure the consistency between the description of each fracture feature from the newly extracted fracture features from multiple periods and that in the atlas, normalize the fracture feature data, and remove spaces, brackets, and other symbols.

[0109] S2: Represent the information such as crack length, width, depth, direction, roughness index, slope and temperature in the crack sample library in the form of word vectors, calculate the cosine similarity of the corresponding word vectors in turn, realize the alignment of crack feature information, and determine whether each crack feature in the newly extracted crack features of multiple periods is already included in the crack sample library.

[0110] S3: Combined with the above similarity judgment, the similarity calculation results are summarized and the position and key feature information of each crack feature in the newly extracted crack features of multiple periods are automatically marked. If the similarity is high, the system automatically marks the new crack feature with higher similarity as an identified crack. Among them, a similarity threshold can be extracted and set. If the similarity between a crack feature in the newly extracted crack features of multiple periods and the crack features already in the crack sample library is greater than the similarity threshold, the crack feature is marked as an identified crack, that is, the crack belongs to the loess landslide crack.

[0111] The characteristics of each crack are identified according to the similarity, and the identification results of crack characteristics in multiple periods are obtained.

[0112] In addition, in one or more embodiments of the present invention, a fissure evolution trend map of a loess landslide is constructed based on the identification results of fissure characteristics in multiple periods, specifically including:

[0113] The data corresponding to the recognition results of the crack features of multiple periods are converted into a spatial data format to obtain the spatial data corresponding to the crack features of multiple periods, and the spatial data corresponding to the crack features of multiple periods are marked with time stamp information.

[0114] The fracture features of multiple periods marked with timestamp information are loaded into the geographic information system for spatial registration, and different colors and transparencies are set for the fracture features of multiple periods marked with timestamp information.

[0115] The crack characteristics of multiple periods are superimposed on the map layout of the geographic information system, and map elements are added to the map layout to obtain the evolution trend map of loess landslide cracks.

[0116] In this embodiment, the data in the fracture sample library is converted into a standard spatial data format (such as Shapefile, GeoJSON, etc.), and timestamp information is annotated according to the shooting time of each phase of multi-source remote sensing image data to facilitate subsequent time series analysis.

[0117] Load all phases of multi-source remote sensing image data into the GIS software, perform spatial registration, and define different styles (such as color, line type, and markers) for each phase of the fracture feature data. For example, set different colors and transparency for fracture features in different time periods to highlight the changing trends of the fractures.

[0118] Compare the fracture characteristics of different periods, conduct spatial analysis on the GIS platform, check whether the spatial location, shape, size and other attributes of the fracture characteristics have changed, and quantify the evolution process of the fracture characteristics.

[0119] Use the layout function of the GIS software to overlay the evolution process of cracks in different periods onto the map layout, add necessary map elements (such as title, scale, legend, etc.), add time labels for the evolution trend of cracks, and finally save the map as an image file or PDF file, ensuring that the exported map resolution is high enough to help understand the evolution process of cracks in the landslide area.

[0120] This embodiment combines the comparative analysis of crack characteristics with multi-phase image data to effectively identify potential landslide risk areas.

[0121] Figure 3 This is a training curve diagram of a crack recognition model provided by the present invention, with the horizontal axis representing the training batch and the vertical axis representing the accuracy.

[0122] Figure 4 This is a partial result diagram of a crack identification method provided by the present invention. Figure 4 (a) Figure 4 (b) Figure 4 (c) and Figure 4 (d) in the figure is a partial result diagram of automatic crack identification. Figure 4 The identification result of (a) is 90% that it is a crack of loess landslide. Figure 4 The identification result of (b) is 99% that it is a crack of loess landslide. Figure 4 The identification result of (c) is 98% that it is a crack of loess landslide. Figure 4 The identification result of (d) is that there is no crack. Figure 4 It can be seen that the present invention can accurately identify the location of the cracks of the loess landslide, thereby achieving accurate prevention of loess landslide disasters.

[0123] The above is a method for constructing a loess landslide fissure evolution trend map provided by one or more embodiments of the present invention. Based on the same idea, the present invention also provides a corresponding device for constructing a loess landslide fissure evolution trend map, such as Figure 6 As shown, including:

[0124] The first acquisition module is used to acquire multi-source remote sensing image data of loess landslides, where the multi-source remote sensing image data includes optical image data, laser point cloud data and surface temperature distribution data.

[0125] The reconstruction module is used to obtain reconstruction data of the loess landslide based on multi-source remote sensing image data, and extract the crack characteristics of the loess landslide based on the reconstruction data.

[0126] The classification module is used to establish a feature extraction atlas library based on multi-source remote sensing image data, and obtain the classification results of the crack features based on the feature extraction atlas library, crack features and position parameters of the crack features.

[0127] The recognition module is used to recognize the crack features collected in multiple periods according to the classification results of the crack features, and obtain recognition results of the crack features in multiple periods.

[0128] The second acquisition module is used to construct a crack evolution trend map of the loess landslide based on the identification results of crack characteristics in multiple periods.

[0129] The specific definition of a device for constructing a loess landslide fissure evolution trend map can be found in the definition of a method for constructing a loess landslide fissure evolution trend map above, and will not be repeated here. The various modules in the device for constructing a loess landslide fissure evolution trend map can be implemented in whole or in part by software, hardware, or a combination thereof. The modules can be embedded in or independent of a processor in a computer device in hardware form, or can be stored in a memory in a computer device in software form, so that the processor can call and execute the operations corresponding to the above modules.

[0130] The present invention also provides a computer-readable storage medium, which stores a computer program, which can be used to execute the Figure 1 A method for constructing a loess landslide crack evolution trend map is provided.

[0131] The present invention also provides Figure 7 The structural diagram of the computer equipment shown in FIG. Figure 7 As shown, at the hardware level, the computer device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory, and may also include other hardware required for the business. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to implement the Figure 1A method for constructing a loess landslide crack evolution trend map is provided.

[0132] Those skilled in the art will appreciate that all or part of the processes in the embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the various methods described. Among them, any reference to memory, storage, database or other media used in the various embodiments provided by the present invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0133] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of the present invention.

Claims

1. A method for constructing a loess landslide fissure evolution trend map, characterized in that: include: Acquiring multi-source remote sensing image data of a loess landslide, wherein the multi-source remote sensing image data includes optical image data, laser point cloud data, and surface temperature distribution data; Acquiring reconstruction data of the loess landslide based on the multi-source remote sensing image data, and extracting crack characteristics of the loess landslide based on the reconstruction data; Establishing a feature extraction atlas library based on the multi-source remote sensing image data, and obtaining a classification result of the crack feature based on the feature extraction atlas library, the crack feature, and the position parameters of the crack feature; Identifying the crack features collected at multiple periods according to the classification results of the crack features to obtain identification results of the crack features at the multiple periods; Converting the data corresponding to the identification results of the crack features of the multiple periods into a spatial data format to obtain spatial data corresponding to the crack features of the multiple periods, and marking the spatial data corresponding to the crack features of the multiple periods with time stamp information; The crack features of multiple periods marked with timestamp information are loaded into a geographic information system for spatial alignment, and the crack features of multiple periods marked with timestamp information are set with different colors and transparencies; the crack features of multiple periods are superimposed on the map layout of the geographic information system, and map elements are added to the map layout to obtain the loess landslide crack evolution trend map.

2. The method for constructing a loess landslide fissure evolution trend map according to claim 1, characterized in that: Obtaining the reconstructed data of the loess landslide according to the multi-source remote sensing image data specifically includes: Acquiring texture information of a fissure region of the loess landslide based on the optical image data, and reconstructing an optical image of the loess landslide according to the texture information; generating a three-dimensional reconstruction model of the ground and exposed areas of the loess landslide based on the laser point cloud data; The thermal response characteristics of the fissure area are reconstructed based on the surface temperature distribution data, and a thermal infrared model of the loess landslide is obtained according to the thermal response characteristics.

3. The method for constructing a loess landslide fissure evolution trend map according to claim 1, wherein: Extracting the fissure characteristics of the loess landslide according to the reconstructed data specifically includes: establishing an image dataset about the fissure based on the reconstructed data; Performing grayscale processing on the image dataset using an edge detection algorithm, and obtaining an edge dataset of the crack based on the grayscale processed image dataset; A segmentation algorithm is used to extract features from the edge data set of the crack to obtain the crack features of the loess landslide.

4. The method for constructing a loess landslide fissure evolution trend map according to claim 1, wherein: Establishing a feature extraction atlas library based on the multi-source remote sensing image data, and obtaining a classification result of the crack feature based on the feature extraction atlas library, the crack feature, and the position parameters of the crack feature, specifically including: Building a fracture sample library based on the multi-source remote sensing image data; Extracting characteristic parameters of each crack in the crack sample library, wherein the characteristic parameters include crack position, crack morphology and temperature; The fractures in the fracture sample library are classified according to their characteristic parameters, and the fractures in the classified fracture sample library are combined with the fracture characteristics of the loess landslide and the position parameters of the loess landslide to obtain a classification result of the fracture characteristics of the loess landslide.

5. The method for constructing a loess landslide fissure evolution trend map according to claim 4, characterized in that: After obtaining the classification results of the fissure characteristics of the loess landslide, the following steps are also included: Obtaining physical parameters of each of the cracks, converting the physical parameters of each of the cracks into standardized values, and obtaining characteristic vectors corresponding to the physical parameters of each of the cracks based on the standardized values, wherein the physical parameters include length, width, depth, direction, roughness, slope, and temperature; The characteristic vectors corresponding to the physical parameters of each crack are standardized and stored using a database system to obtain a standardized crack sample library, and the crack characteristics of the loess landslide are re-identified using the standardized crack sample library to obtain an optimized crack characteristic classification result.

6. The method for constructing a loess landslide fissure evolution trend map according to claim 1, wherein: According to the classification results of the crack features, the crack features collected in multiple periods are identified to obtain the identification results of the crack features in the multiple periods, which specifically include: respectively calculating the similarity between each of the crack features in the plurality of periods and each type of crack feature in the classification results of the crack features; The respective crack features are identified according to the similarity to obtain identification results of the crack features of the multiple periods.

7. A device for constructing a loess landslide fissure evolution trend map, characterized in that: include: A first acquisition module is used to acquire multi-source remote sensing image data of loess landslides, wherein the multi-source remote sensing image data includes optical image data, laser point cloud data and surface temperature distribution data; a reconstruction module, configured to obtain reconstruction data of the loess landslide based on the multi-source remote sensing image data, and extract crack features of the loess landslide based on the reconstruction data; A classification module is used to establish a feature extraction atlas library based on the multi-source remote sensing image data, and obtain a classification result of the crack feature based on the feature extraction atlas library, the crack feature and the position parameter of the crack feature; an identification module, configured to identify the crack features collected at multiple periods according to the classification results of the crack features, and obtain identification results of the crack features at the multiple periods; A second acquisition module is configured to convert the data corresponding to the identification results of the crack features of the multiple periods into a spatial data format, obtain the spatial data corresponding to the crack features of the multiple periods, and mark the spatial data corresponding to the crack features of the multiple periods with time stamp information; The crack features of multiple periods marked with timestamp information are loaded into a geographic information system for spatial alignment, and the crack features of multiple periods marked with timestamp information are set with different colors and transparencies; the crack features of multiple periods are superimposed on the map layout of the geographic information system, and map elements are added to the map layout to obtain the loess landslide crack evolution trend map.

8. A computer-readable storage medium, characterized in that The storage medium stores a computer program, and when the computer program is executed by a processor, the method for constructing a loess landslide fissure evolution trend map according to any one of claims 1 to 6 is implemented.

9. A computer device, characterized in that: The invention comprises a memory, a processor and a computer program stored in the memory and capable of running on the processor. When the processor executes the program, a method for constructing a loess landslide fissure evolution trend map according to any one of claims 1 to 6 is implemented.

Images