Method, system, device and storage medium for distinguishing InSAR deformation types with multi-level constraints

Abstract

The present application relates to a method, system, device and storage medium for distinguishing InSAR deformation types with multi-level constraints. The above method includes: obtaining InSAR annual average deformation phase data and slope data within the scope of the study area; constructing an InSAR deformation area detection network with multi-channel input; using the detection network to obtain a plurality of binary classification maps, determining the boundaries of multiple types of InSAR deformation areas; performing geographic information system spatial overlay analysis to gradually distinguish the types of InSAR deformation areas; obtaining the distribution positions of InSAR deformation area types at the regional scale, obtaining InSAR deformation-derived products, and providing basic data for geological disaster monitoring and early warning and management decision-making.

Description

Technical Field

[0001] This application relates to the technical field of geological disaster monitoring, and particularly to a method, system, and storage medium for distinguishing InSAR deformation types with multi-level constraints. Background Art

[0002] The InSAR technology is one of the important means of ground observation technology and has been widely used in long-term monitoring of small surface deformations. By obtaining surface deformation information through multi-temporal interferometric measurement and mastering the mechanism and trend of surface deformation, it helps to evaluate potential risks. Currently, the InSAR deformation type distinction method mainly uses the persistent scatterer interferometric measurement technology to obtain a discrete point set PS composed of high-coherence pixels, calculates the deformation of it, and separates the active PS and inactive PS points through a clustering method. Combining additional data sources such as landslide catalog data, urban vector data, and terrain data in the study area, it uses expert interpretation to semi-automatically distinguish landslide active deformation areas, ground settlement deformation areas, etc. However, currently, using the persistent scatterer interferometric measurement technology to distinguish deformation area types involves a large amount of data and low operation efficiency, and there are certain limitations for the monitoring application of large-scale and wide-area surface deformations.

[0003] In addition, although the current InSAR technology has played an important role in the applications of monitoring landslides, urban ground settlement, mining monitoring, and glacier movement. However, for the different spatial distribution characteristics of landslides, mining collapses, settlements, and glacier deformations, the degree of automation of InSAR deformation area classification detection is relatively low. Summary of the Invention

[0004] To solve at least one of the above-mentioned technical problems, this application proposes a method, system, and storage medium for distinguishing InSAR deformation types with multi-level constraints.

[0005] According to some embodiments of this application, a method for distinguishing InSAR deformation types with multi-level constraints is provided. The method includes:

[0006] Step 1: Obtain the annual average InSAR deformation phase data and slope data within the scope of the study area; perform band synthesis on the annual average InSAR deformation phase data and slope data to generate multi-source fusion data; wherein, the multi-source fusion data is four-band channel data;

[0007] Step 2: Construct an InSAR deformation area detection network with multi-channel input; use the detection network to perform feature extraction and prediction segmentation on the first three-band channel data of the multi-source fusion data to obtain a first binary classification map;

[0008] Step 3: Assign a projection coordinate system to the first binary classification map and perform raster vectorization processing to obtain the first deformation area vector layer, and determine the boundary of the I - type InSAR deformation area;

[0009] Step 4: Use the detection network to train and predict the data of the four - band channel of the multi - source fusion data to obtain the second binary classification map; assign a projection coordinate system to the second binary classification map and perform raster vectorization processing to obtain the second deformation area vector layer, and determine the boundary of the II - type InSAR deformation area;

[0010] Step 5: Add the sample data of the mining collapse deformation area on the slope as negative samples; repeat Step 4 for training and prediction to obtain the third binary classification map; assign a projection coordinate system to the third binary classification map and perform raster vectorization processing to obtain the third deformation area vector layer, and determine the boundary of the III - type InSAR deformation area;

[0011] Step 6: Perform a geographical information system spatial overlay analysis on the boundary of the III - type InSAR deformation area and the glacier contour vector layer within the study area, and determine that the vector layer that coincides with the glacier contour vector layer within the study area is the glacier deformation area, and the rest is the landslide deformation area on the slope.

[0012] Step 7: Perform a geographical information system spatial overlay analysis on the boundary of the landslide deformation area on the slope and the glacier deformation area determined in Step 6 and the boundary of the II - type InSAR deformation area to determine the mining collapse deformation area on the slope;

[0013] Step 8: Perform a geographical information system spatial overlay analysis on the boundary of the landslide deformation area on the slope and the glacier deformation area determined in Step 6, the boundary of the mining collapse deformation area on the slope determined in Step 7 and the boundary of the I - type InSAR deformation area to determine the settlement deformation area in the plain area.

[0014] In some possible implementation manners, the obtaining of the InSAR annual average deformation phase data and the slope data within the study area includes:

[0015] Use multi - period InSAR data to obtain the InSAR annual average deformation phase data within the study area;

[0016] Crop the range of each InSAR annual average deformation phase data to obtain the corresponding DEM data, and calculate the slope data.

[0017] In some possible implementation manners, the data of the first three band channels of the four - band channel data are multi - period InSAR annual average deformation phase data, and the data of the fourth band channel is slope data.

[0018] In some possible embodiments, the encoding part of the detection network includes a backbone network mobilenetv2 partial structure, an atrous convolution ASPP module, a 1×1 convolution module, and an upsampling module; the decoding part of the detection network includes a fusion module, a 3×3 convolution module, and a 1×1 convolution module.

[0019] In some possible embodiments, the encoding part of the detection network performs deep semantic feature extraction on the first three band-channel data of the multi-source fusion data; after the shallow semantic features obtained by the backbone network mobilenetv2 are convolved by 1×1 in the decoding part of the detection network, they are fused with the deep semantic features obtained by the encoding part, and then abstract semantic features are output after two 3×3 convolution operations, further predicting and segmenting the InSAR deformation area to obtain the first binary classification map.

[0020] In some possible embodiments, the boundaries of the type-I InSAR deformation area include landslide deformation areas on slopes, mining subsidence deformation areas on slopes, settlement deformation areas in plain areas, and glacier deformation areas; the boundaries of the type-II InSAR deformation area include landslide deformation areas on slopes, mining subsidence deformation areas on slopes, and glacier deformation areas; the boundaries of the type-III InSAR deformation area include landslide deformation areas on slopes and glacier deformation areas.

[0021] In some possible embodiments, the glacier contour vector layer within the study area is obtained using the complete catalog vector layer of the global glacier contours published by GLIMS.

[0022] According to some other embodiments of the present application, a multi-level constrained InSAR deformation type discrimination system is provided, and the system includes: a data acquisition module, a data processing module, and a layer analysis module;

[0023] The data acquisition module is used to: obtain the annual average InSAR deformation phase data and slope data within the study area; perform band synthesis on the annual average InSAR deformation phase data and slope data to generate multi-source fusion data; wherein, the multi-source fusion data is four-band channel data;

[0024] The data processing module is used to: construct an InSAR deformation area detection network with multi-channel inputs; use the detection network to perform feature extraction and prediction segmentation on the first three band-channel data of the multi-source fusion data to obtain the first binary classification map;

[0025] Perform projection coordinate system assignment and raster vectorization processing on the first binary classification map to obtain the first deformation area vector layer and determine the boundaries of the type-I InSAR deformation area;

[0026] Training and predicting the data of four band channels of the multi-source fusion data by using the detection network to obtain a second binary classification map; performing projection coordinate system assignment and raster vectorization processing on the second binary classification map to obtain a second deformed area vector layer, and determining the boundary of the II - type InSAR deformed area;

[0027] Increasing the sample data of the mining collapse deformed area on the slope as negative samples; repeating the training and prediction to obtain a third binary classification map; performing projection coordinate system assignment and raster vectorization processing on the third binary classification map to obtain a third deformed area vector layer, and determining the boundary of the III - type InSAR deformed area;

[0028] The layer analysis module is used for: performing a geographic information system spatial overlay analysis on the boundary of the III - type InSAR deformed area and the glacier contour vector layer within the study area, determining that the vector layer coinciding with the glacier contour vector layer within the study area is the glacier deformed area, and the rest is the landslide deformed area on the slope.

[0029] Performing a geographic information system spatial overlay analysis on the determined landslide deformed area boundary and glacier deformed area boundary on the slope and the boundary of the II - type InSAR deformed area to determine the mining collapse deformed area on the slope;

[0030] Performing a geographic information system spatial overlay analysis on the determined landslide deformed area boundary and glacier deformed area boundary on the slope, the mining collapse deformed area boundary on the slope and the boundary of the I - type InSAR deformed area to determine the settlement deformed area in the plain area.

[0031] According to some other embodiments of the present application, there is provided a multi - level constrained InSAR deformation type discrimination device, including a processor and a memory. At least one instruction or at least one program segment is stored in the memory, and the at least one instruction or at least one program segment is loaded and executed by the processor to implement the multi - level constrained InSAR deformation type discrimination method as described above.

[0032] According to some other embodiments of the present application, there is provided a storage medium, in which at least one instruction or at least one program segment is stored, and the at least one instruction or at least one program segment is loaded and executed by the processor to implement the multi - level constrained InSAR deformation type discrimination method as described above.

[0033] Implementing the embodiments of the present application has the following beneficial effects:

[0034] In view of the problem of low automation in the classification and detection of InSAR deformation areas, an embodiment of the present invention provides a method for distinguishing InSAR deformation types. Based on multi-temporal InSAR annual average deformation phase diagrams, aiming at the different spatial distribution characteristics of landslides, mining subsidence, settlement, and glacier deformation, a deep learning method is used, combined with multi-level constraints, to gradually distinguish the types of InSAR deformation areas; obtain the distribution positions of InSAR deformation area types at the regional scale, and obtain InSAR deformation derivative products, providing basic data for geological disaster monitoring and early warning, and management decision-making. The method provided by the present invention is suitable for deformation monitoring applications in research areas with a wide-area scale range, greatly improving the operation efficiency and response speed, and having a high degree of automated classification and detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions and advantages in the embodiments of this specification or the prior art, the following will briefly introduce the drawings required to be used in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of this specification. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0036] Figure 1 The flowchart showing the steps of a method for distinguishing InSAR deformation types with multi-level constraints according to an embodiment of the present application;

[0037] Figure 2 The data processing flowchart showing a method for distinguishing InSAR deformation types with multi-level constraints according to an embodiment of the present application;

[0038] Figure 3 The schematic diagram showing the sample data of the mining subsidence deformation area on a slope according to an embodiment of the present application;

[0039] Figure 4 The structural block diagram showing a system for distinguishing InSAR deformation types with multi-level constraints according to an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0040] The following will clearly and completely describe the technical solutions in the embodiments of this specification with reference to the drawings in the embodiments of this specification. Obviously, the described embodiments are only some embodiments of this specification, rather than all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0041] Combined with Figure 1 and Figure 3 , an embodiment of the present application provides a method for distinguishing InSAR deformation types with multi-level constraints, including:

[0042] S100: Obtain the InSAR annual average deformation phase data and slope data within the study area; perform band synthesis on the InSAR annual average deformation phase data and slope data to generate multi-source fusion data; wherein, the multi-source fusion data is four-band channel data.

[0043] In the embodiments of the present application, the obtaining of the InSAR annual average deformation phase data and slope data within the study area includes:

[0044] Obtain the InSAR annual average deformation phase data within the study area by using multi-period InSAR data;

[0045] Crop the range of each piece of InSAR annual average deformation phase data to obtain the corresponding DEM data, and calculate the slope data.

[0046] The first three band channel data of the four-band channel data are multi-period InSAR annual average deformation phase data, and the fourth band channel data is slope data.

[0047] S200: Construct an InSAR deformation area detection network with multi-channel input; use the detection network to perform feature extraction and prediction segmentation on the first three band channel data of the multi-source fusion data to obtain a first binary classification map.

[0048] Combined with Figure 2 And Tables 1 and 2, the encoding part of the detection network includes a partial structure of the backbone network mobilenetv2, an atrous convolution ASPP module, a 1X1 convolution module, and an upsampling module; the decoding part of the detection network includes a fusion module, a 3X3 convolution module, and a 1X1 convolution module. Among them, the 1X1 convolution module is Conv2d 1×1, the 3X3 convolution module is Conv2d 3×3, and the upsampling module is the F.interpolate function.

[0049] The encoding part of the detection network performs deep semantic feature extraction on the first three band channel data of the multi-source fusion data; the decoding part of the detection network performs 1×1 convolution on the shallow semantic features obtained by the backbone network mobilenetv2, fuses them with the deep semantic features obtained by the encoding part, and then outputs abstract semantic features after two 3×3 convolution operations, further predicting and segmenting the InSAR deformation area to obtain a first binary classification map. In the embodiments of the present application, only a partial structure of the original mobilenetv2 network is used in the deep and shallow semantic feature extraction processes, as shown in Table 1.

[0050] Table 1 Partial structure parameters of the mobilenetv2 backbone network

[0051]

[0052]

[0053] ASPP is a technology for image processing, whose full name is Atrous Spatial Pyramid Pooling, which combines dilated convolution and spatial pyramid pooling processing. The ASPP structure parameters in the embodiments of this application are shown in Table 2.

[0054] The left side of Table 2 encodes the ASPP structure parameters

[0055]

[0056] S300: Perform projection coordinate system assignment and raster vectorization processing on the first binary classification map to obtain a first deformation area vector layer, and determine the boundary of the Class I InSAR deformation area.

[0057] In the embodiments of this application, the boundary of the Class I InSAR deformation area includes the landslide deformation area on the slope, the mining subsidence deformation area on the slope, the settlement deformation area in the plain area, and the glacier deformation area.

[0058] S400: Use the detection network to train and predict the data of the four band channels of the multi-source fusion data to obtain a second binary classification map; perform projection coordinate system assignment and raster vectorization processing on the second binary classification map to obtain a second deformation area vector layer, and determine the boundary of the Class II InSAR deformation area.

[0059] In the embodiments of this application, the boundary of the Class II InSAR deformation area includes the landslide deformation area on the slope, the mining subsidence deformation area on the slope, and the glacier deformation area. Through the terrain features introduced in this step, the settlement deformation area in the plain area is removed.

[0060] S500: Increase the sample data of the mining subsidence deformation area on the slope as negative samples; repeat S400 for training and prediction to obtain a third binary classification map; perform projection coordinate system assignment and raster vectorization processing on the third binary classification map to obtain a third deformation area vector layer, and determine the boundary of the Class III InSAR deformation area. In the embodiments of this application, this step uses the same network as S400 for training and prediction. Through this step, the negative sample knowledge is enhanced, the sampling ratio of positive and negative samples is optimized, and the mining subsidence deformation area on the slope is removed. The boundary of the Class III InSAR deformation area includes the landslide deformation area on the slope and the glacier deformation area. Figure 3 A sample data of the mining subsidence deformation area on the slope in the embodiments of this application is shown. This data is the deformation phase map caused by mining subsidence obtained after processing the InSAR data and is used as a negative sample.

[0061] S600: Perform a geographic information system (GIS) spatial overlay analysis on the boundary of the Class III InSAR deformation area and the glacier contour vector layer within the study area to determine that the vector layer that coincides with the glacier contour vector layer within the study area is the glacier deformation area, and the rest is the landslide deformation area on the slope. In the embodiments of the present application, the glacier contour vector layer within the study area is obtained using the complete catalog vector layer of the global glacier contours published by GLIMS.

[0062] S700: Perform a GIS spatial overlay analysis on the boundary of the landslide deformation area on the slope and the glacier deformation area determined in S600 and the boundary of the Class II InSAR deformation area to determine the mining collapse deformation area on the slope.

[0063] S800: Perform a GIS spatial overlay analysis on the boundary of the landslide deformation area on the slope and the glacier deformation area determined in S600, and on the boundary of the mining collapse deformation area on the slope determined in S700 and the boundary of the Class I InSAR deformation area to determine the settlement deformation area in the plain area.

[0064] In the embodiments of the present application, through the above multi-level steps of constraint, the InSAR deformation types can be quickly distinguished.

[0065] The method embodiments provided in the embodiments of the present application can be executed in electronic devices such as mobile terminals, computer terminals, servers, or similar computing devices.

[0066] The above embodiments have detailedly introduced the method for distinguishing InSAR deformation types with multi-level constraints. Implementing the embodiments of the present application has the following beneficial effects:

[0067] Based on multi-temporal InSAR annual average deformation phase diagrams, for the different spatial distribution characteristics of landslides, mining collapses, settlements, and glacier deformations, using deep learning methods and combining multi-level constraints, gradually distinguish the types of InSAR deformation areas; obtain the distribution locations of InSAR deformation area types at the regional scale, obtain InSAR deformation derivative products, and provide basic data for geological disaster monitoring and early warning, and management decision-making. The method provided by the present invention is suitable for deformation monitoring applications in study areas with a wide-area scale range, greatly improves the computing efficiency and response speed, and has a high degree of automated classification detection.

[0068] The embodiments of the present application also provide a multi-level constraint InSAR deformation type distinguishing system 10. Specifically, please refer to Figure 4 The system 10 includes: a data acquisition module 100, a data processing module 200, and a layer analysis module 300;

[0069] The data acquisition module 100 is used to: obtain the InSAR annual average deformation phase data and slope data within the study area; perform band synthesis on the InSAR annual average deformation phase data and slope data to generate multi-source fusion data; wherein, the multi-source fusion data is four-band channel data.

[0070] The data processing module 200 is used to: construct an InSAR deformation area detection network with multi-channel input; use the detection network to perform feature extraction and prediction segmentation on the first three band channel data of the multi-source fusion data to obtain a first binary classification map.

[0071] Perform projection coordinate system assignment and raster vectorization processing on the first binary classification map to obtain a first deformation area vector layer and determine the boundary of the Class I InSAR deformation area.

[0072] Use the detection network to train and predict the four band channel data of the multi-source fusion data to obtain a second binary classification map; perform projection coordinate system assignment and raster vectorization processing on the second binary classification map to obtain a second deformation area vector layer and determine the boundary of the Class II InSAR deformation area.

[0073] Add sample data of mining subsidence deformation areas on slopes as negative samples; repeat training and prediction to obtain a third binary classification map; perform projection coordinate system assignment and raster vectorization processing on the third binary classification map to obtain a third deformation area vector layer and determine the boundary of the Class III InSAR deformation area.

[0074] The layer analysis module 300 is used to: perform geographic information system spatial overlay analysis on the boundary of the Class III InSAR deformation area and the glacier contour vector layer within the study area, and determine that the vector layer that coincides with the glacier contour vector layer within the study area is the glacier deformation area, and the rest is the landslide deformation area on the slope.

[0075] Perform geographic information system spatial overlay analysis on the determined landslide deformation area boundary and glacier deformation area boundary on the slope and the boundary of the Class II InSAR deformation area to determine the mining subsidence deformation area on the slope.

[0076] Perform geographic information system spatial overlay analysis on the determined landslide deformation area boundary and glacier deformation area boundary on the slope, the mining subsidence deformation area boundary on the slope and the boundary of the Class I InSAR deformation area to determine the settlement deformation area in the plain area.

[0077] In the embodiment of the present application, the data acquisition module 100, the data processing module 200, and the layer analysis module 300 can also be used for more specific processing operations corresponding to the respective steps of the above method.

[0078] An embodiment of the present application further provides a multi-level constrained InSAR deformation type discrimination device, including a processor and a memory. At least one instruction or at least one program segment is stored in the memory, and the at least one instruction or at least one program segment is loaded and executed by the processor to implement the above-mentioned multi-level constrained InSAR deformation type discrimination method. Among them, the memory can be used to store software programs and modules. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory. In addition, the memory may include a high-speed random access memory and may also include a non-volatile memory. Correspondingly, the memory may further include a memory controller to provide the processor with access to the memory. Optionally, the device may further include one or more CPUs or multiple GPUs, one or more power supplies, one or more wired or wireless network interfaces, one or more input / output interfaces 940, and / or one or more operating systems, etc.

[0079] An embodiment of the present application further provides a storage medium. At least one instruction or at least one program segment is stored in the storage medium, and the at least one instruction or at least one program segment is loaded and executed by the processor to implement the above-mentioned multi-level constrained InSAR deformation type discrimination method.

[0080] The embodiments of the present application have been described above. The above description is exemplary and not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations are obvious to those of ordinary skill in the art in the technical field without departing from the scope and spirit of the described embodiments. The choice of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to technologies in the market, or to enable other ordinary skill in the art in the technical field to understand the disclosed embodiments.

Claims

1. A method for distinguishing InSAR deformation types with multi-level constraints, characterized in that The method includes: Step 1: Obtain the InSAR annual average deformation phase data and slope data within the study area; perform band synthesis on the InSAR annual average deformation phase data and slope data to generate multi-source fusion data; wherein, the multi-source fusion data is four-band channel data; Step 2: Construct an InSAR deformation area detection network with multi-channel input; use the detection network to perform feature extraction and predictive segmentation on the first three band channel data of the multi-source fusion data to obtain a first binary classification map; Step 3: Perform projection coordinate system assignment and raster vectorization processing on the first binary classification map to obtain a first deformation area vector layer, and determine the boundary of the Class I InSAR deformation area; Step 4: Use the detection network to train and predict the four band channel data of the multi-source fusion data to obtain a second binary classification map; perform projection coordinate system assignment and raster vectorization processing on the second binary classification map to obtain a second deformation area vector layer, and determine the boundary of the Class II InSAR deformation area; Step 5: Add sample data of mining subsidence deformation areas on slopes as negative samples; repeat Step 4 for training and prediction to obtain a third binary classification map; perform projection coordinate system assignment and raster vectorization processing on the third binary classification map to obtain a third deformation area vector layer, and determine the boundary of the Class III InSAR deformation area; Step 6: Perform geographic information system spatial overlay analysis on the boundary of the Class III InSAR deformation area and the glacier contour vector layer within the study area, and determine that the vector layer that coincides with the glacier contour vector layer within the study area is the glacier deformation area, and the rest are landslide deformation areas on slopes; Step 7: Perform geographic information system spatial overlay analysis on the boundary of the landslide deformation area on slopes and the glacier deformation area determined in Step 6 and the boundary of the Class II InSAR deformation area to determine the mining subsidence deformation area on slopes; Step 8: Perform geographic information system spatial overlay analysis on the boundary of the landslide deformation area on slopes and the glacier deformation area determined in Step 6, the boundary of the mining subsidence deformation area on slopes determined in Step 7 and the boundary of the Class I InSAR deformation area to determine the settlement deformation area in the plain area.

2. The method according to claim 1, wherein The obtaining of the InSAR annual average deformation phase data and slope data within the study area includes: Obtain the InSAR annual average deformation phase data within the study area using multi-period InSAR data; Crop the range of each InSAR annual average deformation phase data to obtain the corresponding DEM data, and calculate the slope data.

3. The method according to claim 1, wherein The first three band channel data of the four-band channel data are multi-period InSAR annual average deformation phase data, and the fourth band channel data is slope data.

4. The method according to claim 1, characterized in that, Wherein, The encoding part of the detection network includes the backbone network mobilenetv2 partial structure, the atrous convolution ASPP module, the 1X1 convolution module and the upsampling module; the decoding part of the detection network includes the fusion module, the 3X3 convolution module and the 1X1 convolution module.

5. The method according to claim 4, wherein The encoding part of the detection network extracts deep semantic features from the data of the first three band channels of the multi-source fusion data; the decoding part of the detection network performs 1×1 convolution on the shallow semantic features obtained by the backbone network mobilenetv2, fuses them with the deep semantic features obtained by the encoding part, and then outputs abstract semantic features through two 3×3 convolution operations, further predicting and segmenting the InSAR deformation area to obtain the first and second classification maps.

6. The method according to claim 1, wherein The boundaries of the Class I InSAR deformation areas include landslide deformation areas on slopes, mining subsidence deformation areas on slopes, settlement deformation areas in plain areas, and glacier deformation areas; the boundaries of the Class II InSAR deformation areas include landslide deformation areas on slopes, mining subsidence deformation areas on slopes, and glacier deformation areas; the boundaries of the Class III InSAR deformation areas include landslide deformation areas on slopes and glacier deformation areas.

7. The method according to claim 1, characterized in that The glacier contour vector layer within the study area is obtained using the complete catalog vector layer of the global glacier contours published by GLIMS.

8. A multi-level constrained InSAR deformation type discrimination system, characterized in that, The system includes: a data acquisition module, a data processing module, and a layer analysis module; The data acquisition module is used to: obtain the annual average InSAR deformation phase data and slope data within the study area; perform band synthesis on the annual average InSAR deformation phase data and slope data to generate multi-source fusion data; wherein, the multi-source fusion data is four-band channel data; The data processing module is used to: construct an InSAR deformation area detection network with multi-channel inputs; use the detection network to perform feature extraction and prediction segmentation on the data of the first three band channels of the multi-source fusion data to obtain the first and second classification maps; Perform projection coordinate system assignment and raster vectorization processing on the first and second classification maps to obtain the first deformation area vector layer and determine the boundaries of the Class I InSAR deformation areas; Use the detection network to train and predict the data of the four band channels of the multi-source fusion data to obtain the second and second classification maps; perform projection coordinate system assignment and raster vectorization processing on the second and second classification maps to obtain the second deformation area vector layer and determine the boundaries of the Class II InSAR deformation areas; Add sample data of mining subsidence deformation areas on slopes as negative samples; repeat training and prediction to obtain the third and second classification maps; perform projection coordinate system assignment and raster vectorization processing on the third and second classification maps to obtain the third deformation area vector layer and determine the boundaries of the Class III InSAR deformation areas; The layer analysis module is used to: perform geographic information system spatial overlay analysis on the boundaries of the Class III InSAR deformation areas and the glacier contour vector layer within the study area, determine that the vector layer that coincides with the glacier contour vector layer within the study area is the glacier deformation area, and the rest are landslide deformation areas on slopes; Perform geographic information system spatial overlay analysis on the determined boundaries of the landslide deformation areas on slopes and the glacier deformation areas and the boundaries of the Class II InSAR deformation areas to determine the mining subsidence deformation areas on slopes; Perform a geographic information system spatial overlay analysis on the determined boundaries of the landslide deformation area and the glacier deformation area on the slope, the mining subsidence deformation area boundary on the slope, and the Class I InSAR deformation area boundary to determine the settlement deformation area in the plain area.

9. An InSAR deformation type discrimination device with multi-level constraints, characterized in that It includes a processor and a memory. At least one instruction or at least one program segment is stored in the memory, and at least one instruction or at least one program segment is loaded and executed by the processor to implement the method for distinguishing InSAR deformation types with multi-level constraints as described in any one of the preceding claims 1-7.

10. A storage medium, characterized in that, At least one instruction or at least one program segment is stored in the storage medium, and at least one instruction or at least one program segment is loaded and executed by the processor to implement the method for distinguishing InSAR deformation types with multi-level constraints as described in any one of the preceding claims 1-7.

Images