Full-polarization SAR image landslide detection method

Abstract

The invention, which relates to the technical field of landslide detection, discloses a fully-polarized SAR image landslide detection method comprising the following steps: collecting a single fully-polarized SAR image of a landslide target area and carrying out data preprocessing; based on a data preprocessing result, taking the depolarization degree as a constraint, and combining double-sided angle scattering and surface scattering to construct a landslide identification detector; using a generalized gamma distribution model to represent statistical information of the landslide identification detector; and de-noising the binary image obtained by the represented landslide identification detector by using a morphological operator to obtain a landslide detection result, thereby completing the fully polarized SAR image landslide detection. According to the method, the technical bottleneck that the landslide position distribution is difficult to quickly and accurately obtain based on a single full-polarization SAR image in the prior art is solved.

Description

Technical Field

[0001] The invention relates to the technical field of landslide detection, and in particular to a landslide detection method using full-polarization SAR images. Background Art

[0002] Landslides usually occur after major triggering events such as earthquakes or heavy rainfall. Accurate and timely landslide mapping is of great significance for post-disaster rescue, disaster assessment, and the protection of personal and property safety. To achieve this goal, high-resolution optical images have become an important means to extract landslide information after a disaster. However, high-resolution optical images are often obscured by clouds and fog, making it difficult to provide reliable information quickly after a disaster occurs. Synthetic aperture radar (SAR) has all-weather and all-day earth observation capabilities and has gradually become a mainstream tool for landslide identification research. Existing landslide identification methods based on SAR images usually rely on multiple SAR images as input and require a large number of ground truth samples for training. This feature limits its application in scenarios where data is scarce or timeliness is high. In addition, most studies have not fully explored the potential of single-phase SAR images to extract landslide information.

[0003] Fully polarized SAR images have shown advantages in landslide mapping. By alternately sending and receiving microwaves with horizontal (H) and vertical (V) polarization, they make up for the lack of information in single-polarized SAR data. However, most of the current studies based on polarized SAR images use methods such as change detection and coherence analysis, which rely on high-resolution data before the event, such as optical images, SAR images, and digital elevation models (DEMs) before the event. However, high-resolution data before the event is not easy to obtain, which has become a major obstacle to timely acquisition of landslide distribution information. A few studies have attempted to use supervised classification algorithms to process fully polarized SAR images to extract landslide distribution information. This method can, to a certain extent, use a single fully polarized SAR image for landslide identification. However, it requires a large amount of sample point data for training, and the model training process is time-consuming, which cannot quickly respond to post-disaster mapping needs. Therefore, there is still much room for improvement in the existing technology in terms of efficient, accurate, and independent of pre-event data landslide mapping capabilities. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the present invention provides a landslide detection method using full-polarization SAR images to solve the problems existing in the prior art.

[0005] In order to achieve the above-mentioned invention object, the technical solution adopted by the present invention is: a landslide detection method using full polarization SAR images, comprising the following steps: S1: Collect a single fully polarized SAR image of the landslide target area and perform data preprocessing; S2: Based on the data preprocessing results, the depolarization degree is used as a constraint and combined with dihedral angle scattering and surface scattering to construct a landslide identification detector; S3: Characterizing the statistical information of landslide identification detectors using the generalized gamma distribution model; S4: Use morphological operators to denoise the binary image obtained by the characterized landslide identification detector to obtain the landslide detection result and complete the landslide detection of full polarization SAR images.

[0006] Furthermore, the data preprocessing in S1 includes multi-view processing, geocoding operation and initial denoising processing; The multi-view processing is to reduce random noise and improve the signal-to-noise ratio of the SAR image by superimposing multiple views of the SAR image. for:

[0007] in, is the number of views in the azimuth direction, is the number of views in the distance direction, Azimuth index and distance index The corresponding pixel value; The geocoding operation is: converting the SAR image from the range-azimuth coordinate system to the geographic coordinate system for subsequent geographic positioning and analysis of the landslide area; The initial denoising process is to use a non-local filter to remove speckle noise from a single fully polarized SAR image of the landslide target area.

[0008] Furthermore, the landslide identification detector in S2 is:

[0009]

[0010] in, represents the landslide identification detector, is the dihedral angular scattering, is surface scattering, is the degree of depolarization, is the maximum value of the Stokes vector, , , and is the Stokes vector.

[0011] Furthermore, the probability density function of the generalized gamma distribution model in S3 is for:

[0012] in, is the power parameter, is the shape parameter, is the scale parameter, is the variable in the generalized gamma distribution model, is the Gamma function.

[0013] Furthermore, the detection threshold of the generalized gamma distribution model in S3 is for:

[0014] in, is the inverse incomplete Gamma function, is a given false alarm probability value.

[0015] Furthermore, the morphological operator in S4 for:

[0016] in, For the opening operation, It is a closing operation.

[0017] The beneficial effects of the present invention are: 1) the difference between the landslide target and its background is enhanced by using mathematical formulas, which significantly improves the distinguishability of landslide features and thus enhances the recognition effect; 2) by combining scattering mechanism analysis with polarization anisotropy characteristics, the present invention can effectively overcome the technical bottleneck of the prior art that it is difficult to quickly and accurately obtain the landslide location distribution based on a single fully polarized SAR image; 3) it can meet the rapid response requirements in landslide disaster monitoring and identification, and provide real-time support for disaster emergency response. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 This is a flow chart of a landslide detection method using full-polarization SAR images.

[0019] Figure 2 This is a structural block diagram of a landslide detection method using full-polarization SAR images.

[0020] Figure 3 This is the landslide detection result diagram of this embodiment. DETAILED DESCRIPTION

[0021] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0022] like Figure 1 and Figure 2 As shown, a landslide detection method using full polarization SAR images comprises the following steps: S1: Collect a single fully polarized SAR image of the landslide target area and perform data preprocessing; S2: Based on the data preprocessing results, the depolarization degree is used as a constraint and combined with dihedral angle scattering and surface scattering to construct a landslide identification detector; S3: Characterizing the statistical information of landslide identification detectors using the generalized gamma distribution model; S4: Use morphological operators to denoise the binary image obtained by the characterized landslide identification detector to obtain the landslide detection result and complete the landslide detection of full polarization SAR images.

[0023] The basic idea of ​​this invention is to capture the performance of landslide targets in terms of scattering mechanism and polarization anisotropy characteristics based on the polarization characteristics of landslides in full-polarimetric SAR images, and to enhance the significant difference between the target and the background by using mathematical basic formulas. ) can achieve accurate acquisition of landslide targets, thus solving the problem that landslide targets are difficult to capture quickly in full-polarization SAR images.

[0024] The data preprocessing in S1 includes multi-view processing, geocoding operation and initial denoising; The multi-view processing is to reduce random noise and improve the signal-to-noise ratio of the SAR image by superimposing multiple views of the SAR image. for:

[0025] in, is the number of views in the azimuth direction, is the number of views in the distance direction, Azimuth index and distance index The corresponding pixel value; The geocoding operation is: converting the SAR image from the range-azimuth coordinate system to the geographic coordinate system for subsequent geographic positioning and analysis of the landslide area; The initial denoising process is to use a non-local filter to remove speckle noise from a single fully polarized SAR image of the landslide target area.

[0026] In this embodiment, multi-view processing is first performed according to the characteristics of the fully polarized SAR image to reduce random noise in the image and improve the signal-to-noise ratio; geocoding operation is performed to make the image data accurately correspond to the actual geographical status of the ground; finally, a non-local filter is used to denoise the image, and the filter window is set to 11×11 to further improve the image quality by removing speckle noise.

[0027] Based on the relatively strong scattering characteristics of landslides in full-polarization SAR images, the present invention introduces the depolarization degree ( ) as a constraint, combined with bihedral angular scattering and surface scattering , construct a landslide identification detector.

[0028] The landslide identification detector in S2 is:

[0029]

[0030] in, represents the landslide identification detector, is the dihedral angular scattering, is surface scattering, is the degree of depolarization, is the maximum value of the Stokes vector, , , and is the Stokes vector.

[0031] The normalized value of is between 0 and 1. When the polarized wave interacts with the random medium, the polarization state of the backscattered wave changes, and the purity of the scattered wave The change of the polarization of the transmitted wave causes the change of the wave entropy. This method overcomes the interference of strong scatterers such as buildings, reduces the computational complexity, and effectively improves the efficiency of the algorithm.

[0032] After defining a new detector, the statistical information of the detector must be mapped to characterize the adaptive detection of landslides in different scenarios. To characterize the statistical behavior of the landslide detector.

[0033] Probability density function of the generalized gamma distribution model in S3 for:

[0034] in, is the power parameter, is the shape parameter, is the scale parameter, is the variable in the generalized gamma distribution model, is the Gamma function.

[0035] The detection threshold of the generalized gamma distribution model in S3 for:

[0036] in, is the inverse incomplete Gamma function, is a given false alarm probability value.

[0037] The above steps effectively utilize The detection framework is used to detect landslides on full polarimetric SAR images, but the binary images obtained by landslide detectors are often distorted by salt and pepper noise, especially in full polarimetric SAR images, where speckle noise is independently distributed in the image and causes significant changes in reflectivity. The present invention adopts morphological image processing to eliminate defects by considering the form and structure of the image.

[0038] That is, the introduction of morphological operators (i.e., opening operations) and closed operation ), to solve problems such as internal holes and blurred boundaries, and overcome the limitations of fixed structural elements at the edge of the target area. In binary images, the opening operation removes small white objects (such as salt and pepper noise) from the black background, while the closing operation fills small black holes in the white target area or connects isolated points to make its edges smooth.

[0039] The morphological operators in S4 for:

[0040] in, For the opening operation, It is a closing operation.

[0041] These operations will not significantly change the area of ​​the target area, but can effectively reduce noise and produce better landslide detection output. Through the implementation of the above content, the accurate extraction of landslide targets can be completed efficiently based on full-polarization SAR images, breaking through the technical bottleneck of existing polarization SAR landslide identification methods. This method not only improves the accuracy of landslide identification, but also significantly accelerates the response speed of landslide detection, thus forming a technical approach for rapid landslide identification based on full-polarization SAR images, providing a reliable and efficient solution, and providing strong technical support for landslide disaster monitoring and emergency response.

[0042] In one embodiment of the present invention, Figure 3 As shown in the figure, the landslide area is detected by the method of the present invention. The white area in the figure represents the detected landslide area, while the black area represents the non-landslide area. It can be seen that the method can accurately distinguish the landslide area from the background area and has a high recognition accuracy, which fully reflects the effectiveness and applicability of the method in landslide detection.

[0043] The present invention is particularly aimed at the technical difficulties that landslide targets are widely distributed, complex in shape and difficult to effectively identify in actual scenes. The scattering mechanism and polarization anisotropy characteristics of landslide targets in full-polarization SAR images are fully considered, and a basic mathematical formula is constructed for feature calculation. Finally, a generalized gamma distribution model ( ) to accurately obtain the landslide target. This method does not need to introduce machine learning technology to complete the acquisition of sample data and long-term training, thereby further realizing the rapid and automatic identification of landslide targets.

[0044] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific variations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed by the present invention, and these variations and combinations are still within the protection scope of the invention.

Claims

1. A landslide detection method using full polarization SAR images, characterized in that: The following steps are involved: S1: Collect a single fully polarized SAR image of the landslide target area and perform data preprocessing; S2: Based on the data preprocessing results, the depolarization degree is used as a constraint and combined with dihedral angle scattering and surface scattering to construct a landslide identification detector; S3: Characterizing the statistical information of landslide identification detectors using the generalized gamma distribution model; S4: Use morphological operators to denoise the binary image obtained by the characterized landslide identification detector to obtain the landslide detection result and complete the landslide detection of full polarization SAR images.

2. The method for landslide detection using full polarization SAR images according to claim 1, characterized in that: The data preprocessing in S1 includes multi-view processing, geocoding operation and initial denoising; The multi-view processing is to reduce random noise and improve the signal-to-noise ratio of the SAR image by superimposing multiple views of the SAR image. for: in, is the number of views in the azimuth direction, is the number of views in the distance direction, Azimuth index and distance index The corresponding pixel value; The geocoding operation is: converting the SAR image from the range-azimuth coordinate system to the geographic coordinate system for subsequent geographic positioning and analysis of the landslide area; The initial denoising process is to use a non-local filter to remove speckle noise from a single fully polarized SAR image of the landslide target area.

3. The method for landslide detection using full polarization SAR images according to claim 1, characterized in that: The landslide identification detector in S2 is: in, represents the landslide identification detector, is the dihedral angular scattering, is surface scattering, is the degree of depolarization, is the maximum value of the Stokes vector, , , and is the Stokes vector.

4. The method for landslide detection using full polarization SAR images according to claim 1, characterized in that: Probability density function of the generalized gamma distribution model in S3 for: in, is the power parameter, is the shape parameter, is the scale parameter, is the variable in the generalized gamma distribution model, is the Gamma function.

5. The method for landslide detection using full polarization SAR images according to claim 4 is characterized in that: The detection threshold of the generalized gamma distribution model in S3 for: in, is the inverse incomplete Gamma function, is a given false alarm probability value.

6. The method for landslide detection using full polarization SAR images according to claim 1, characterized in that: The morphological operators in S4 for: in, For the opening operation, It is a closing operation.

Images