Method and device for detecting coherence change of fused terrain mountainous area sar image

Abstract

This invention discloses a method for detecting coherence changes in SAR images of mountainous areas by incorporating terrain features. It also discloses a computer device. The method obtains the coherence change results before and after an event based on an ascending orbit image set and the coherence change results before and after an event based on a descending orbit image set. Then, it calculates the pixel compression coefficients of the ascending orbit image set and the normalized pixel compression coefficients of the descending orbit image set based on terrain elevation data. Finally, it uses the two normalized pixel compression coefficients as weighting coefficients for the coherence change results before and after an event in the corresponding image sets, obtaining the coherence change detection results incorporating terrain factors. This invention enhances the applicability and stability of the method under complex mountainous conditions by introducing terrain factors. Compared with existing methods, it overcomes the interference of geometric distortions such as perspective contraction, overlay, and shadows caused by a single orbital viewpoint, avoiding the problems of unclear or false detection of coherence changes.

Description

Technical Field

[0001] This invention belongs to the field of remote sensing disaster detection technology, specifically involving a method for detecting coherence changes in mountainous SAR images with fused terrain, and also involving computer equipment. Based on ascending and descending orbit SAR images, it utilizes R-Index weighting and fusion of terrain factors to detect coherence changes in complex mountainous areas, thereby achieving the purpose of detecting surface disturbances. Background Technology

[0002] In environments where traditional ground-based observation conditions are limited, spaceborne SAR (Synthetic Aperture Radar), as an important active remote sensing technology, has become a crucial tool for studying surface changes caused by sudden events due to its advantages such as wide coverage, fast response speed, non-contact operation, and insensitivity to weather and lighting conditions. SAR imagery can not only acquire millimeter-level deformation information through interferometric phase analysis, but its coherence index can also sensitively reflect surface disturbances and damage. Especially in scenarios with abrupt changes in surface structure, coherence change detection can be used to quickly identify affected areas. However, due to the side-looking imaging of SAR, the ground monitoring visibility of a single orbit in areas with large topographic relief will be affected by geometric distortion, resulting in unclear coherence change detection results or even false detections. Effectively fusing complementary information from ascending and descending orbits and overcoming their respective perspective limitations is a key challenge.

[0003] Geometric distortion is caused by the side-looking imaging mode of SAR satellites and is related to the observation geometry between the radar line of sight and the local terrain. It is mainly divided into overlay and shadow. According to the relative geometry between the satellite and the ground, when the landslide faces the satellite, if the incident angle is greater than the slope angle, the length of the slope projected onto the SAR image is compressed, resulting in perspective contraction. If the incident angle is less than the slope angle, the top and bottom of the slope in the SAR image will be inverted, resulting in active overlay. At the same time, the surrounding area will also be passively overlaid due to the influence of active overlay. Usually, the overlaid area has more scattered echoes mixed in, which appears brighter in the SAR image. When the landslide faces away from the satellite and the incident angle is greater than the complementary angle of the slope angle, the satellite cannot illuminate the back slope area, resulting in active shadow. At the same time, because the SAR satellite cannot obtain signals due to obstacles between it and the ground surface, passive shadow will be formed, which will appear as black in the SAR image.

[0004] In existing technologies, the area of ​​SAR image overlay and shadow regions is calculated based on the geometric relationship between the radar line of sight and local terrain. An early example is the Layover Shadow Map (LSM) method, which is based on DEM and sensor acquisition parameters. It simulates the active and passive areas of geometric distortion based on satellite position, satellite azimuth, and terrain information. However, the LSM method is relatively complex to calculate and has low efficiency in obtaining geometric distortion. Subsequently, the improved LSM method incorporates the influence of local terrain into the slant range index (R-Index) and has been tested, obtaining good results. Through continuous improvement, the R-Index method has been gradually developed, which can more accurately detect and identify the range of geometric distortion and judge the quality of SAR images.

[0005] In complex mountainous areas, monitoring along a single orbital direction is susceptible to geometric distortion, making it difficult to obtain complete and reliable coherence change detection results. In contrast, combining the perspectives of ascending and descending orbits can significantly improve regional visibility. R-Index represents the ratio of the pixel size of the tilt to the pixel size of the geometric shape, also known as the pixel compression coefficient. The degree of pixel compression is related to the slope incidence angle; a larger R-Index indicates better image quality presented by SAR images in the monitored area. Based on ascending and descending orbit R-Index weighting, interference caused by geometric distortions such as perspective contraction, overlay, and shadows from a single orbit can be effectively overcome in complex mountainous areas, avoiding unclear or even false detection results of coherence changes. This is of great significance for surface disturbance detection in complex mountainous areas. Summary of the Invention

[0006] The purpose of this invention is to address the aforementioned problems in the existing technology by providing a method for detecting coherence changes in mountainous SAR images that incorporates terrain. It also provides computer equipment that, by fusing multi-view information from the ascending and descending orbits and introducing terrain constraints, can effectively reduce interference caused by geometric distortions such as perspective contraction, overlay, and shadows in single-orbit SAR images, improve the identifiability of disturbance responses, and reduce the risk of missed and false detections.

[0007] The above-mentioned objectives of the present invention are achieved by the following technical means: A method for detecting coherence changes in mountainous SAR images with topographical integration includes the following steps: Step 1: Obtain the original ascending orbit image set and the original descending orbit image set of the event location before and after the event, and perform preprocessing on them respectively to obtain the ascending orbit image set and the descending orbit image set. Step 2: Construct pre-event and post-event ascent interferometric pairs in the ascent image set, and calculate the coherence change results of the ascent image set before and after the event based on the pre-event and post-event ascent interferometric pairs. Pre-event and post-event deorbiting interferometric pairs were constructed in the deorbiting image set. Based on the pre-event and post-event deorbiting interferometric pairs, the coherence changes of the deorbiting image set before and after the event were calculated. Step 3: Calculate the pixel compression coefficients of the ascending orbit image set and the descending orbit image set based on the terrain elevation data, and then normalize them. Step 4: Using the pixel compression coefficients of the normalized ascending orbit image set and the normalized descending orbit image set as weight coefficients, perform weighted summation on the coherence change results before and after the event for both the ascending orbit image set and the descending orbit image set to obtain the coherence change detection results fused with terrain factors.

[0008] The coherence change detection results, which incorporate topographic factors as described above, are calculated based on the following formula: ; In the formula, To integrate the coherence change detection results of topographic factors, The pixel compression factor for the normalized up-track image set. For the pixel compression factor of the normalized down-track image set, This represents the results of changes in coherence before and after events in the up-orbit image set. This represents the results of coherence changes before and after events in the de-orbiting image set.

[0009] As mentioned above, both the original ascending orbit image set and the original descending orbit image set are SAR image sets. The original ascending orbit image set includes two ascending orbit SAR images before the event and one ascending orbit SAR image after the event. The original descending orbit image set includes two descending orbit SAR images before the event and one descending orbit SAR image after the event.

[0010] As described above, the preprocessing process includes track correction, geocoding, registration, and deskewing, specifically comprising the following steps: Step 1.2: Obtain the POD precise orbit data for the corresponding time period of the original ascending orbit image set and the POD precise orbit data for the corresponding time period of the original descending orbit image set, and obtain the terrain elevation data covering the event occurrence area based on the latitude and longitude range provided by the parameter file of the SAR image. Step 1.3: Perform track correction on the original track of the original ascending track image set based on the POD precise track data of the corresponding time period; perform track correction on the original track of the original descending track image set based on the POD precise track data of the corresponding time period. Step 1.4: Select the SAR image with the middle time in the original ascending image set as the main image of the original ascending image set, and generate a mapping lookup table between the main image of the original ascending image set after orbit correction and the terrain elevation data. Based on the mapping lookup table between the main image of the original ascending image set after orbit correction and the terrain elevation data, geocode all SAR images in the original ascending image set after orbit correction. The SAR image with the middle time in the original down-orbit image set is selected as the main image of the original down-orbit image set, and a mapping lookup table between the main image of the original down-orbit image set after orbit correction and the terrain elevation data is generated. Based on the mapping lookup table between the main image of the original down-orbit image set after orbit correction and the terrain elevation data, all SAR images in the original down-orbit image set after orbit correction are geocoded. Step 1.5: Register other SAR images in the geocoded original ascending orbit image set to the main image, and register other SAR images in the geocoded original descending orbit image set to the main image, to obtain the registered original ascending orbit image set and the registered original descending orbit image set respectively. Step 1.6: Perform deskewing processing on the registered original up-track image set and the registered original down-track image set to obtain the up-track image set and the down-track image set.

[0011] The interferometric pair includes a primary image and a secondary image; wherein, the secondary image of the pre-event ascending orbit interferometric pair is the ascending orbit SAR image before the event occurs; the secondary image of the post-event ascending orbit interferometric pair is the ascending orbit SAR image after the event occurs; the secondary image of the pre-event descending orbit interferometric pair is the descending orbit SAR image before the event occurs; and the secondary image of the post-event descending orbit interferometric pair is the descending orbit SAR image after the event occurs. The pre-event coherence of the ascending orbit image set is calculated based on the pre-event ascending orbit interferometry pair, the post-event coherence of the ascending orbit image set is calculated based on the post-event ascending orbit interferometry pair, and the change in coherence of the ascending orbit image set before and after the event is calculated based on the pre-event and post-event coherence of the ascending orbit image set. The pre-event coherence of the down-orbit image set is calculated based on the pre-event down-orbit interferometry pair, and the post-event coherence of the down-orbit image set is calculated based on the post-event down-orbit interferometry pair. Then, the change in coherence of the down-orbit image set before and after the event is calculated based on the pre-event and post-event coherence of the down-orbit image set.

[0012] Coherence is calculated based on the following formula: ; In the formula, For coherence, This is the size of the local window used for coherence calculations. and The main image and the auxiliary image are respectively in the first... Complex amplitude at each pixel for conjugate, for conjugate, For the first The differential interference phase of each pixel, For the first in the local window Pixel weights; The imaginary unit; For modulo operation; The change in coherence before and after an event is the coherence after the event minus the coherence before the event.

[0013] As described above, step 3 specifically includes the following steps: Step 3.1: Calculate the local slope angle and local aspect angle of the area where the event occurred; Step 3.2: Calculate the angle of incidence and heading of the ascending orbit image set and the angle of incidence and heading of the descending orbit image set in the area where the event occurred; Step 3.3: Calculate the pixel compression coefficient of the up-track image set based on the following formula: ; In the formula, is the pixel compression factor for the upscaling image set. The angle of incidence for the ascending orbit image set. For the heading angle of the ascending image set, For local slope angle, This refers to the local slope angle; The pixel compression factor of the downtracked image set is calculated based on the following formula: ; In the formula, is the pixel compression factor for the down-track image set. The incident angle of the down-orbit image set, The heading angle for the descending orbit image set; Step 3.4: Normalize the pixel compression coefficients of the up-track image set and the down-track image set. The normalization is based on the following formula: ; .

[0014] As described above, the local slope angle and the local aspect angle are calculated based on the following formulas: ; ; ; In the formula, for Increment in direction for Increment in direction This represents the percentage of slope.

[0015] A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the method for detecting coherence changes in mountainous SAR images with fused terrain as described above.

[0016] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method for detecting coherence changes in mountainous SAR images with fused terrain as described above.

[0017] Compared with the prior art, the present invention has the following advantages: (1) Universality: The method of the present invention can effectively identify surface disturbances by utilizing coherence changes; by introducing topographic factors, the applicability and stability of the method under complex mountain conditions are enhanced, so that it can be widely and flexibly applied to the detection of surface disturbances in complex mountain areas.

[0018] (2) Accuracy: Compared with existing methods, the method of the present invention overcomes the interference of geometric distortions such as perspective shrinkage, overlay and shadow caused by a single orbital viewpoint, avoids the problem of insignificant or false detection of coherence changes, and thus more accurately identifies the spatial distribution characteristics of surface disturbances in complex mountainous areas.

[0019] (3) Practicality: The method of the present invention can efficiently and accurately detect surface disturbances in complex mountainous areas, and is suitable for large-scale surface change surveys in complex mountainous areas. It can also provide technical support for applications such as geological disaster monitoring and building damage assessment, and has strong engineering application potential and promotion value.

[0020] Compared with existing detection methods, this invention addresses SAR geometric distortion in complex mountainous areas by integrating topographic factors and multi-view information from the ascent and descent orbits to improve the accuracy of identifying the spatial distribution of surface disturbances. Furthermore, it enhances the stability and reliability of the results through automated fine registration, differential interferometry, filtering, and quality-weighted fusion processes, achieving more efficient and comprehensive disturbance detection. Attached Figure Description

[0021] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a diagram showing the coherence changes before and after an event in the Sentinel-1 ASAR image set of region A in Embodiment 2 of the present invention. The solid yellow boxes represent areas with good visibility (F2 and F3), and the dashed yellow boxes represent areas of perspective contraction (F1). Figure 3 This is a diagram showing the coherence changes before and after an event in the Sentinel-1 ASAR image set of region A in Embodiment 2 of the present invention. The solid yellow box represents the region with good visibility (F1), and the dashed yellow box represents the regions with perspective contraction (F2 and F3). Figure 4 This is a diagram showing the coherence change detection results of fused terrain factors in region A of Embodiment 2 of the present invention, where the yellow solid boxes represent areas with good visibility (F1, F2, and F3). Figure 5 For the present invention Figure 2 Optical image of the F1 region prior to the event; Figure 6 For the present invention Figure 2 Optical image of the F2 region prior to the event; Figure 7 For the present invention Figure 2 Optical image of the central F3 region prior to the event; Figure 8 For the present invention Figure 2 Optical image of the area F1 following the incident; Figure 9 For the present invention Figure 2 Optical image of the region F2 following the event; Figure 10 For the present invention Figure 2 Optical image of the area in the middle F3 region after the event; Figure 11 This is a diagram showing the coherence changes before and after an event in the Sentinel-1 ASAR image set of region B in Embodiment 2 of the present invention. The yellow solid boxes represent areas with good visibility (I1, I2, and I3), and the black boxes represent overlapping shadow areas (I4). Figure 12 This is a diagram showing the coherence change results of the Sentinel-1 ASAR image set in region B of Embodiment 2 of the present invention before and after an event. The black areas represent the overlay shadow regions (I1, I2, I3, I4, I5, I6, I7, I8, and I9). Figure 13 This is a diagram showing the coherence change detection results of the fused terrain factors in region B of Embodiment 2 of the present invention, where the yellow solid boxes represent areas with good visibility (I1, I2, and I3). Figure 14 For the present invention Figure 11 Optical image of the I1 region prior to the event; Figure 15 For the present invention Figure 11 Optical image of the I2 region prior to the event; Figure 16 For the present invention Figure 11 Optical image of the I3 region prior to the event; Figure 17 For the present invention Figure 11 Optical image of the I1 region following the event; Figure 18 For the present invention Figure 11 Optical image of the I2 region following the event; Figure 19 For the present invention Figure 11 Optical image of the I3 region after the event. Detailed Implementation

[0022] To facilitate understanding and implementation of the present invention by those skilled in the art, the present invention will be further described in detail below with reference to embodiments. The embodiments described herein are for illustration and explanation only and are not intended to limit the present invention. Example 1:

[0023] like Figure 1 As shown, the method for detecting coherence changes in mountainous SAR images with fused terrain includes the following steps: Step 1: Obtain the original ascending and descending image sets of the event location before and after the event, and preprocess them separately to obtain the ascending and descending image sets. This includes the following steps: Step 1.1: Use the ASF website (Alaska Satellite Facility, one of the world's leading free SAR image distribution platforms) to delineate the area where the incident occurred, select the original ascending orbit image set and the original descending orbit image set before and after the incident based on the time of the incident, and use the ASF website to automatically download them.

[0024] In this embodiment, the original ascending orbit image set includes two ascending orbit SAR images before the event and one ascending orbit SAR image after the event; the original descending orbit image set includes two descending orbit SAR images before the event and one descending orbit SAR image after the event.

[0025] Step 1.2: Use the open-source tool Sentineleof (a command-line tool specifically designed for automatically and in batches downloading Sentinel satellite precise orbit ephemeris data) to automatically download the POD precise orbit data for the corresponding time period of the original ascending orbit image set and the POD precise orbit data for the corresponding time period of the original descending orbit image set. Then, use the elevation tool to automatically obtain the topographic elevation data (DEM) of the area where the event occurred based on the latitude and longitude range provided in the parameter file of the SAR image.

[0026] Step 1.3: Use GAMMA software (a synthetic aperture radar interferometry (InSAR) software) to perform orbit correction on the original orbit of the original ascending orbit image set based on the POD precise orbit data of the corresponding time period; and perform orbit correction on the original orbit of the original descending orbit image set based on the POD precise orbit data of the corresponding time period; in order to reduce or even eliminate the impact of orbit errors on subsequent registration and differential interferometry processing.

[0027] Step 1.4: Select the SAR image with the middle time in the original ascending image set as the main image of the original ascending image set. Use GAMMA software to generate a mapping lookup table between the main image of the original ascending image set after orbit correction and the terrain elevation data. Based on the mapping lookup table between the main image of the original ascending image set after orbit correction and the terrain elevation data, geocode all SAR images in the original ascending image set after orbit correction.

[0028] The SAR image with the middle time in the original down-orbit image set is selected as the main image of the original down-orbit image set. The GAMMA software is used to generate a mapping lookup table between the main image of the original down-orbit image set after orbit correction and the terrain elevation data. Based on the mapping lookup table between the main image of the original down-orbit image set after orbit correction and the terrain elevation data, all SAR images in the original down-orbit image set after orbit correction are geocoded.

[0029] Step 1.5: Use GAMMA software to register other SAR images in the geocoded original up-orbit image set to the main image, and register other SAR images in the geocoded original down-orbit image set to the main image, so that the registration offset is less than 0.0005 pixels and the weight value of each burst (strip) is not 0, to obtain the registered original up-orbit image set and the registered original down-orbit image set respectively.

[0030] Step 1.6: Use GAMMA software to perform deskewing on the registered original up-orbit image set and the registered original down-orbit image set to reduce the biased estimation caused by phase gradient changes, and complete the preprocessing to obtain the up-orbit image set and the down-orbit image set.

[0031] Step 2: Construct pre-event and post-event ascending orbit interferometric pairs in the ascending orbit image set, and calculate the coherence change results of the ascending orbit image set before and after the event based on the pre-event and post-event ascending orbit interferometric pairs; construct pre-event and post-event descending orbit interferometric pairs in the descending orbit image set, and calculate the coherence change results of the descending orbit image set before and after the event based on the pre-event and post-event descending orbit interferometric pairs. Specifically, this includes the following steps: Step 2.1: Construct two sets of interferometric pairs in the ascending orbit image set and the descending orbit image set respectively. The interferometric pair includes a main image and a secondary image. Specifically: the secondary image of the pre-event ascending orbit interferometric pair is the ascending orbit SAR image before the event; the secondary image of the post-event ascending orbit interferometric pair is the ascending orbit SAR image after the event; the secondary image of the pre-event descending orbit interferometric pair is the descending orbit SAR image before the event; and the secondary image of the post-event descending orbit interferometric pair is the descending orbit SAR image after the event.

[0032] Step 2.2: Use GAMMA software to perform batch differential interferometry and adaptive filtering on the four sets of interferometric pairs constructed in Step 2.1 to obtain the temporal coherence of the ascending and descending image sets.

[0033] Coherence is calculated based on the following formula: (1); In the formula, To calculate the coherence, This represents the size of the local window used for coherence calculation. and The main image and the auxiliary image are respectively in the first... Complex amplitude at each pixel for conjugate, for conjugate, For the first The differential interference phase of each pixel; For the first in the local window The pixel weights are calculated using a Gaussian function based on the distance between pixels; The imaginary unit; This is a modulo operation.

[0034] The pre-event coherence of the ascending orbit image set is calculated using formula (1) based on the pre-event ascending orbit interferometer, and the post-event coherence of the ascending orbit image set is calculated based on the post-event ascending orbit interferometer; the pre-event coherence of the descending orbit image set is calculated using formula (1) based on the pre-event descending orbit interferometer, and the post-event coherence of the descending orbit image set is calculated based on the post-event descending orbit interferometer.

[0035] Step 2.3: On the Anaconda Spyder compilation platform, calculate the change in coherence of the image set before and after the event based on the pre-event coherence and post-event coherence of the image set. Denote the change in coherence of the image set before and after the event as follows: The changes in coherence before and after events in the down-orbit image set are calculated based on the pre-event and post-event coherence. These changes are denoted as follows: By using the ascending and descending image sets to calculate the corresponding changes in coherence before and after the event, the interference caused by factors such as satellite positioning errors, vegetation changes, and random noise can be reduced.

[0036] Anaconda is a data science platform and package manager based on the Python language, while Spyder is an integrated development environment designed for scientific computing and data analysis.

[0037] The change in coherence before and after the event is the coherence after the event minus the coherence before the event. The value of the change in coherence is in the range of -1 to 1. If the change in coherence before and after the event is less than 0, it means that the coherence of the interference pair after the event is reduced compared to the interference pair before the event.

[0038] Step 3: Extract local slope angle and local aspect angle based on terrain elevation data, calculate the pixel compression coefficient (R-Index) of the ascending orbit image set and the pixel compression coefficient of the descending orbit image set respectively, and normalize them to numerically evaluate the imaging quality of ascending and descending orbit SAR images in complex mountainous areas.

[0039] Step 3.1: Based on topographic elevation data, use ArcGIS (a comprehensive Geographic Information System (GIS) platform) to calculate the local slope angle and local aspect angle of the land surface in the event area, based on the following formulas: (2); (3); (4); In the formula, It represents the percentage of slope. For local slope angle, This is the local slope angle. for Increment in direction (elevation direction), for Increment in direction (horizontal direction).

[0040] Step 3.2: Use GAMMA software to calculate the incident angle and heading angle of the ascending orbit image set and the descending orbit image set in the event occurrence area.

[0041] Since the incident angle and heading angle of images in the same area and on the same track (ascending or descending) are the same, each image can be selected for calculation.

[0042] Step 3.3: Calculate the pixel compression coefficients of the up-orbit image set and the down-orbit image set using Anaconda's Spyder compilation platform, based on the following formula: (5); (6); In the formula, is the pixel compression factor for the upscaling image set. The angle of incidence for the ascending orbit image set. For the heading angle of the ascending image set, is the pixel compression factor for the down-track image set. The incident angle of the down-orbit image set, The heading angle for the descending orbit image set.

[0043] Step 3.4: Normalize the pixel compression coefficients of the up-track image set and the down-track image set using Anaconda's Spyder compilation platform, based on the following formula: (7); (8); In the formula, The pixel compression factor for the normalized up-track image set. R-Index is the pixel compression factor for the normalized down-track image set, and the pixel compression factor R-Index for the overlapping shadow region is set to 0.

[0044] Step 4: Using Anaconda's Spyder compilation platform, the pixel compression coefficients of the normalized ascending orbit image set and the normalized descending orbit image set are used as weighting coefficients. The coherence change results before and after the event in both the ascending and descending orbit image sets are weighted and summed to obtain the coherence change detection results fused with terrain factors, based on the following formula: (9); In the formula, To integrate the coherence change detection results of topographic factors, This represents the results of changes in coherence before and after events in the up-orbit image set. The results show the changes in coherence before and after the event in the down-orbit image set. The greater the decrease in coherence, the more severe the surface damage caused by the event.

[0045] Finally, high-resolution optical images before and after the event were acquired, and cross-validation was performed on each region with decreased coherence.

[0046] This invention targets complex mountainous areas where ascending-orbit SAR images and descending-orbit SAR images are complementary in geometric characteristics. By weighting the ascending-orbit and descending-orbit R-Index and combining the coherence change detection results of the ascending-orbit and descending-orbit images, the invention can effectively avoid the situation where the detection of coherence changes is not obvious or even erroneous due to geometric distortion in mountainous areas. Example 2:

[0047] This invention selects regions A and B, both located in mountainous areas with complex terrain, where a sudden event caused significant damage to buildings and changes in the landform. To detect the surface disturbance caused by this event in the complex mountainous area, the method of this invention is used to analyze regions A and B to identify the building damage and changes in landform caused by the sudden event.

[0048] Step 1 of Implementation Example 1: Select the original ascending-orbit Sentinel-1A SAR image set (i.e., the original ascending-orbit SAR image set of the Sentinel-1A satellite) and the original descending-orbit Sentinel-1A SAR image set (i.e., the original descending-orbit SAR image set of the Sentinel-1A satellite) covering regions A and B before and after the event, and combine them with the contemporaneous POD precise orbit data to eliminate orbital errors; take the image of the middle period as the main image, perform geocoding by combining it with terrain elevation data, then register the remaining images with the main image, and perform deskewing processing on the registered images to reduce the estimation bias caused by terrain phase gradient, thus obtaining the ascending-orbit Sentinel-1A SAR image set and the descending-orbit Sentinel-1A SAR image set for regions A and B.

[0049] Step 2 of Example 1 is performed: the images are combined into interferometric pairs in chronological order, and batch differential interferometry and adaptive filtering are performed using GAMMA software to extract the coherence change results before and after the event for the ascending Sentinel-1A SAR image set in region A and region B, and the coherence change results before and after the event for the descending Sentinel-1A SAR image set. Then, the surface coherence change map is obtained by performing differential operations on the interferometric pairs.

[0050] Step 3 of Example 1 is executed: the local slope angle and local aspect angle of the monitoring area are extracted from the DEM, and the incident angle and heading angle of the SAR image are combined to calculate the pixel compression coefficient R-Index and normalize it as a weight coefficient. The coherence change results before and after the event of the ascending orbit Sentinel-1A SAR image set and the coherence change results before and after the event of the descending orbit Sentinel-1A SAR image set are weighted and summed to obtain the coherence change detection results of the fused terrain factors in regions A and B.

[0051] Figures 2-10 and Figures 11-19The surface disturbance detection results for regions A and B are presented respectively. It can be seen that the coherence change detection method fused with topographic factors proposed in this invention has significantly improved the detection effect compared with existing methods. It can effectively avoid the situation of unclear detection results or misjudgment, and is highly consistent with the change information of high-resolution optical images. This invention effectively reduces the interference caused by geometric distortions such as perspective contraction, overlay, and shadows in single-orbit SAR images by combining R-Index with multi-view information of ascending and descending orbit SAR images, improves the identifiability of surface disturbance response, and reduces the risk of missed detection and false detection. It is of great significance for surface disturbance detection in complex mountainous areas.

[0052] It should be noted that the embodiments described in this invention are merely illustrative of the spirit of the invention. Those skilled in the art to which this invention pertains can make various modifications or additions to the described embodiments or use similar methods to substitute them, without departing from the spirit of the invention or exceeding the scope defined by the appended claims.

Claims

1. A method for detecting coherence changes in mountainous SAR images fused with terrain features, characterized in that, Includes the following steps: Step 1: Obtain the original ascending orbit image set and the original descending orbit image set of the event location before and after the event, and perform preprocessing on them respectively to obtain the ascending orbit image set and the descending orbit image set. Step 2: Construct pre-event and post-event ascent interferometric pairs in the ascent image set, and calculate the coherence change results of the ascent image set before and after the event based on the pre-event and post-event ascent interferometric pairs. Pre-event and post-event deorbiting interferometric pairs were constructed in the deorbiting image set. Based on the pre-event and post-event deorbiting interferometric pairs, the coherence changes of the deorbiting image set before and after the event were calculated. Step 3: Calculate the pixel compression coefficients of the ascending orbit image set and the descending orbit image set based on the terrain elevation data, and then normalize them. Step 4: Using the pixel compression coefficients of the normalized ascending orbit image set and the normalized descending orbit image set as weight coefficients, perform weighted summation on the coherence change results before and after the event for both the ascending orbit image set and the descending orbit image set to obtain the coherence change detection results fused with terrain factors.

2. The method for detecting coherence changes in mountainous SAR images with fused terrain according to claim 1, characterized in that, The coherence change detection results of the fused terrain factors are calculated based on the following formula: ； In the formula, To integrate the coherence change detection results of topographic factors, For the pixel compression factor of the normalized up-track image set, For the pixel compression factor of the normalized down-track image set, The results show the changes in coherence before and after events in the up-orbit image set. This represents the results of coherence changes before and after events in the de-orbiting image set.

3. The method for detecting coherence changes in mountainous SAR images with fused terrain according to claim 2, characterized in that, Both the original ascending orbit image set and the original descending orbit image set are SAR image sets. The original ascending orbit image set includes two ascending orbit SAR images before the event and one ascending orbit SAR image after the event. The original descending orbit image set includes two descending orbit SAR images before the event and one descending orbit SAR image after the event.

4. The method for detecting coherence changes in mountainous SAR images with fused terrain according to claim 3, characterized in that, The preprocessing includes track correction, geocoding, registration, and deskewing, specifically including the following steps: Step 1.2: Obtain the POD precise orbit data for the corresponding time period of the original ascending orbit image set and the POD precise orbit data for the corresponding time period of the original descending orbit image set, and obtain the terrain elevation data covering the event occurrence area based on the latitude and longitude range provided by the parameter file of the SAR image. Step 1.3: Perform track correction on the original track of the original ascending track image set based on the POD precise track data of the corresponding time period; perform track correction on the original track of the original descending track image set based on the POD precise track data of the corresponding time period. Step 1.4: Select the SAR image with the middle time in the original ascending image set as the main image of the original ascending image set, and generate a mapping lookup table between the main image of the original ascending image set after orbit correction and the terrain elevation data. Based on the mapping lookup table between the main image of the original ascending image set after orbit correction and the terrain elevation data, geocode all SAR images in the original ascending image set after orbit correction. The SAR image with the middle time in the original down-orbit image set is selected as the main image of the original down-orbit image set, and a mapping lookup table between the main image of the original down-orbit image set after orbit correction and the terrain elevation data is generated. Based on the mapping lookup table between the main image of the original down-orbit image set after orbit correction and the terrain elevation data, all SAR images in the original down-orbit image set after orbit correction are geocoded. Step 1.5: Register other SAR images in the geocoded original ascending orbit image set to the main image, and register other SAR images in the geocoded original descending orbit image set to the main image, to obtain the registered original ascending orbit image set and the registered original descending orbit image set respectively. Step 1.6: Perform deskewing processing on the registered original up-track image set and the registered original down-track image set to obtain the up-track image set and the down-track image set.

5. The method for detecting coherence changes in mountainous SAR images with fused terrain according to claim 4, characterized in that, The interferometric pair includes a primary image and a secondary image; wherein, the secondary image of the pre-event ascending orbit interferometric pair is the ascending orbit SAR image before the event occurs; the secondary image of the post-event ascending orbit interferometric pair is the ascending orbit SAR image after the event occurs; the secondary image of the pre-event descending orbit interferometric pair is the descending orbit SAR image before the event occurs; and the secondary image of the post-event descending orbit interferometric pair is the descending orbit SAR image after the event occurs. The pre-event coherence of the ascending orbit image set is calculated based on the pre-event ascending orbit interferometry pair, the post-event coherence of the ascending orbit image set is calculated based on the post-event ascending orbit interferometry pair, and the change in coherence of the ascending orbit image set before and after the event is calculated based on the pre-event and post-event coherence of the ascending orbit image set. The pre-event coherence of the down-orbit image set is calculated based on the pre-event down-orbit interferometry pair, and the post-event coherence of the down-orbit image set is calculated based on the post-event down-orbit interferometry pair. Then, the change in coherence of the down-orbit image set before and after the event is calculated based on the pre-event and post-event coherence of the down-orbit image set.

6. The method for detecting coherence changes in mountainous SAR images with fused terrain according to claim 5, characterized in that, Coherence is calculated based on the following formula: ； In the formula, For coherence, This is the size of the local window used for coherence calculations. and The main image and the auxiliary image are respectively in the first... Complex amplitude at each pixel for conjugate, for conjugate, For the first The differential interference phase of each pixel, For the first in the local window Pixel weights; The imaginary unit; For modulo operation; The change in coherence before and after an event is the coherence after the event minus the coherence before the event.

7. The method for detecting coherence changes in mountainous SAR images with fused terrain according to claim 6, characterized in that, Step 3 specifically includes the following steps: Step 3.1: Calculate the local slope angle and local aspect angle of the area where the event occurred; Step 3.2: Calculate the angle of incidence and heading of the ascending orbit image set and the angle of incidence and heading of the descending orbit image set in the area where the event occurred; Step 3.3: Calculate the pixel compression coefficient of the up-track image set based on the following formula: ； In the formula, is the pixel compression factor for the upscaling image set. The angle of incidence for the ascending image set. For the heading angle of the ascending image set, For local slope angle, This refers to the local slope angle; The pixel compression factor of the downtracked image set is calculated based on the following formula: ； In the formula, is the pixel compression factor for the down-track image set. The incident angle of the down-orbit image set, The heading angle for the descending orbit image set; Step 3.4: Normalize the pixel compression coefficients of the up-track image set and the down-track image set. The normalization is based on the following formula: ； 。 8. The method for detecting coherence changes in mountainous SAR images with fused terrain according to claim 7, characterized in that, The local slope angle and the local aspect angle are calculated based on the following formulas: ； ； ； In the formula, for Increment in direction for Increment in direction This represents the percentage of slope.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method for detecting coherence changes in mountainous SAR images with fused terrain as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for detecting coherence changes in mountainous SAR images with fused terrain as described in any one of claims 1 to 8.

Images