Sar tomography reference net generation method and device based on minimum spanning tree

Abstract

The application relates to a minimum spanning tree-based SAR tomography reference net generation method and device. The method comprises the following steps: calculating amplitude deviation index values of all PS points in a remote sensing imaging area constituting an imaging reference net, selecting a PS point with the minimum amplitude deviation index value from all PS points as a starting reference point according to the amplitude deviation index values of all PS points, adding the starting reference point into an SPS point set, and initializing an SPS edge set as an empty set; the imaging reference net comprises a PS point set; according to a pre-set distance threshold, PS edges meeting the distance threshold are sequentially found in the SPS point set and the PS point set, and a candidate SPS edge set is set; the highest-quality edge is selected from the candidate SPS edge set and added into the SPS edge set, and the end points corresponding to the highest-quality edge are moved from the PS point set to the SPS point set, and finally, an SPS edge set and an SPS point set are output; and a SAR tomography reference net is constructed according to the final SPS edge set and the SPS point set. The method can generate a reference net which maximally covers a full scene.

Description

Technical Field

[0001] This application relates to the field of remote sensing imaging technology, and in particular to a method and apparatus for generating SAR tomographic reference networks based on minimum spanning tree. Background Technology

[0002] Spaceborne SAR (Synthetic Aperture Radar) tomography utilizes multi-angle observations of the same target to form a synthetic aperture in the altitude direction, achieving true 3D imaging of the observed scene. Ideally, multi-baseline SAR data used for SAR tomography is acquired by a multi-baseline radar system through a single flyby (multi-baseline single flyby). This data acquisition method eliminates the influence of atmospheric decoherence and atmospheric phase errors between multi-baseline SAR data. However, the number of array antennas in a typical multi-baseline SAR system is very limited, resulting in a small amount of SAR data obtained through a single flyby and a short synthetic aperture length in the altitude direction, making it difficult to meet the high-quality imaging requirements of SAR tomography. Furthermore, multi-baseline SAR systems are difficult and costly to design, and their large-scale application is still some time away. Currently, the most economical and efficient method for acquiring SAR tomography data is the multi-flyby method. Unlike single-flyby SAR systems, multi-flyby SAR systems inevitably suffer from atmospheric decoherence and atmospheric phase errors between SAR data acquired during each flyby due to the different atmospheric phases experienced at each flyby. For multi-pass SAR systems, the atmospheric phase screen (APS) can be considered a multiplicative noise error. Without correction, it will severely impact SAR tomography, making it difficult to focus the SAR tomographic signal at altitude. Therefore, atmospheric phase correction is essential before SAR tomography; it is a prerequisite. Furthermore, high APS correction accuracy is required to achieve SAR tomography of buildings in urban areas. Because atmospheric phase error exhibits strong temporal randomness with observation time, it is difficult to accurately correct APS error using analytical models. For small areas, the spatial low-pass characteristic of APS can be utilized by selecting one APS point within the area as a reference point and performing phase difference calculations between the remaining APS points and the reference point to significantly offset the APS error. For large areas, selecting a single reference APS point is insufficient for large-area APS correction. At this point, a reference network needs to be established by selecting reference PS points that uniformly cover the entire scene to achieve large-area APS correction. The PS points surrounding the reference network are then connected to the nearest reference PS point to form a star-shaped network. Phase difference analysis is performed on the endpoints of each edge of the star-shaped network to achieve whole-scene APS correction. This is the basic principle of large-area atmospheric phase correction based on a two-layer network. However, APS correction based on a two-layer network still faces the problem of isolated SPS points and islands in reference network creation. Furthermore, high-quality connecting edges are crucial for successful tomographic imaging. Therefore, creating a maximally connected reference network that maximizes coverage of the entire scene and ensures high-quality edge connections is key to successful spaceborne SAR tomographic atmospheric phase correction across the entire scene based on a two-layer network. Summary of the Invention

[0003] Therefore, it is necessary to provide a method and apparatus for generating SAR tomographic reference networks based on minimum spanning tree that can cover the entire scenario to the greatest extent, in order to address the above-mentioned technical problems.

[0004] A method for generating SAR tomographic reference networks based on minimum spanning tree, the method comprising:

[0005] The remote sensing imaging area is acquired, and the amplitude deviation index values ​​of all PS points in the imaging reference network constituting the remote sensing imaging area are calculated. Based on the amplitude deviation index values ​​of all PS points, the PS point with the smallest amplitude deviation index value is selected as the starting reference point, and the starting reference point is added to the SPS point set. The SPS edge set is initialized to an empty set. The imaging reference network includes the PS point set. An SPS point is a point that contains only one permanent scatterer in a single pixel.

[0006] Based on a pre-set distance threshold, sequentially search for PS edges in the SPS point set and PS point set that satisfy the distance threshold, and construct a candidate PS edge set using the PS edges that satisfy the distance threshold; perform phase difference on each edge in the candidate PS edge set and then use the spectral estimation algorithm to perform imaging, and then use the scattering point detection algorithm to detect the imaging results, and use the detection results to set the candidate SPS edge set;

[0007] Select the highest quality edge from the candidate SPS edge set and add it to the SPS edge set. Move the endpoint corresponding to the highest quality edge from the PS point set to the SPS point set to obtain the updated SPS edge set and SPS point set. Stop updating the SPS edge set and SPS point set when the PS point set is empty. Output the final SPS edge set and SPS point set. Construct the SAR tomographic reference network based on the final SPS edge set and SPS point set.

[0008] In one embodiment, the amplitude deviation index value of all PS points constituting the reference net is calculated, including:

[0009] The amplitude deviation index of all PS points constituting the reference net is calculated as follows:

[0010]

[0011] in, Indicates the number of observations. Represents a pixel The The amplitude value of the second observation.

[0012] In one embodiment, the preset distance threshold is set within a range of [range]. In one embodiment, based on a pre-set distance threshold, PS edges that satisfy the distance threshold are sequentially searched in the SPS point set and the PS point set, and a candidate PS edge set is constructed using the PS edges that satisfy the distance threshold, including:

[0013] Based on a pre-set distance threshold, sequentially search for PS edges in the SPS point set and PS point set that satisfy the distance threshold, and construct a candidate PS edge set using the PS edges that satisfy the distance threshold:

[0014]

[0015] in, For SPS point set The point in the middle, PS point set The point in the middle, , These are the spatial coordinates of the reference network vertex and the candidate vertex, respectively. This represents the Euclidean distance operator. In one embodiment, the imaging results are detected according to a scattering point detection algorithm, and a candidate SPS edge set is set using the detection results, including:

[0016] The imaging results are detected using a scattering point detection algorithm. If both endpoints of the PS edge are SPS points and the similarity parameter of the two endpoints of the PS edge is not less than a preset similarity threshold, the PS edge is retained to construct a candidate SPS edge set; otherwise, the PS edge is deleted.

[0017] In one embodiment, the edge with the highest quality is:

[0018]

[0019] in, express Similarity parameters between them This represents the preset similarity threshold. This represents the endpoint corresponding to the edge with the highest quality.

[0020] In one embodiment, the endpoints corresponding to the highest quality edges are moved from the PS point set to the SPS point set, resulting in updated SPS edge sets and SPS point sets, including:

[0021] Move the endpoints corresponding to the highest quality edges from the PS point set to the SPS point set to obtain the updated SPS edge set and SPS point set as follows:

[0022]

[0023] in, .

[0024] A SAR tomographic reference network generation device based on minimum spanning tree, the device comprising:

[0025] The initialization module is used to acquire the remote sensing imaging area, calculate the amplitude deviation index values ​​of all PS points in the imaging reference network constituting the remote sensing imaging area, select the PS point with the smallest amplitude deviation index value from all PS points as the starting reference point, and add the starting reference point to the SPS point set. The SPS edge set is initialized to an empty set. The imaging reference network includes the PS point set. An SPS point is a point that contains only one permanent scatterer in a single pixel.

[0026] The candidate SPS edge set setting module is used to sequentially find PS edges in the SPS point set and PS point set that meet the distance threshold according to the pre-set distance threshold, and construct the candidate PS edge set using the PS edges that meet the distance threshold; after performing phase difference on each edge in the candidate PS edge set, the spectral estimation algorithm is used for imaging, and the imaging result is detected according to the scattering point detection algorithm, and the candidate SPS edge set is set using the detection result.

[0027] The SAR tomographic reference network construction module is used to select the highest quality edge from the candidate SPS edge set and add it to the SPS edge set. It also moves the endpoint corresponding to the highest quality edge from the PS point set to the SPS point set, obtaining the updated SPS edge set and SPS point set. When the PS point set is empty, the update of the SPS edge set and SPS point set is stopped, and the final SPS edge set and SPS point set are output. The SAR tomographic reference network is constructed based on the final SPS edge set and SPS point set.

[0028] The aforementioned SAR tomographic reference network generation method and apparatus based on minimum spanning tree, in this application, selects the starting reference point of the SAR tomographic reference network by calculating the amplitude deviation index values ​​of all PS points in the imaging reference network constituting the remote sensing imaging area. Then, by setting PS edge search and SPS point judgment rules, it sequentially searches for PS edges in the SPS point set and PS point set that satisfy the distance threshold according to a pre-set distance threshold, and constructs a candidate PS edge set using the PS edges that satisfy the distance threshold. After performing phase difference on each edge in the candidate PS edge set, it performs imaging using a spectral estimation algorithm, and then detects the imaging results using a scattering point detection algorithm. The detection results are used to set a candidate SPS edge set, and finally, the candidate SPS edge set is generated from the candidate SPS... The edge with the highest quality is selected from the edge set and added to the SPS edge set. The endpoint corresponding to the edge with the highest quality is moved from the PS point set to the SPS point set, resulting in the updated SPS edge set and SPS point set. The update of the SPS edge set and SPS point set is stopped when the PS point set is empty. By continuously adding new SPS points and SPS edges to the SPS point set and SPS edge set, only one SPS edge is added each time an SPS point is added, and this edge has the highest imaging quality among all candidate SPS edges. Based on the final SPS edge set and SPS point set, the points and edges connected from the starting reference point are all optimal. Constructing a SAR tomographic reference network using the optimal points and edges can guarantee the maximum network coverage and the optimal imaging quality. Attached Figure Description

[0029] Figure 1 This is a flowchart illustrating a SAR tomographic reference network generation method based on minimum spanning tree in one embodiment;

[0030] Figure 2 This is a schematic diagram of the reference network construction in one embodiment;

[0031] Figure 3 This is a schematic diagram of an average SAR intensity map in one embodiment;

[0032] Figure 4 A schematic diagram of a reference net established in another embodiment;

[0033] Figure 5 This is a height map of PS points identified by the method proposed in this application in one embodiment;

[0034] Figure 6 This is a structural block diagram of a SAR tomographic reference network generation device based on minimum spanning tree in one embodiment. Detailed Implementation

[0035] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0036] In one embodiment, such as Figure 1 As shown, a method for generating SAR tomographic reference networks based on minimum spanning trees is provided, including the following steps:

[0037] Step 102: Obtain the remote sensing imaging area, calculate the amplitude deviation index value of all PS points in the imaging reference network constituting the remote sensing imaging area, select the PS point with the smallest amplitude deviation index value from all PS points as the starting reference point, and add the starting reference point to the SPS point set. The SPS edge set is initialized to an empty set. The imaging reference network includes the PS point set. An SPS point is a point that contains only one permanent scatterer in a single pixel.

[0038] Since all PS points in the reference network are SPS points, the initial reference point is also an SPS point. An SPS point is a PS point that contains only one strong scattering source within a single pixel. The candidate PS point with the smallest amplitude deviation index is most likely to be a true SPS point; therefore, the PS point with the smallest amplitude deviation index is chosen as the initial reference point. After the initial SPS point is determined, it is added to the SPS point set. In this case, the SPS point set only contains the initial SPS point, and the SPS edge set is also... Initialize to an empty set. Assume there are a total of If there are 1 PS candidate point, then the PS point set is... There is also One PS point.

[0039] Step 104: Based on the pre-set distance threshold, sequentially search for PS edges in the SPS point set and PS point set that satisfy the distance threshold, and construct a candidate PS edge set using the PS edges that satisfy the distance threshold; perform phase difference on each edge in the candidate PS edge set and then use the spectral estimation algorithm to perform imaging, and then use the scattering point detection algorithm to detect the imaging results, and use the detection results to set the candidate SPS edge set.

[0040] This application limits the length of the SPS edge by setting a distance threshold, so that the APS between the two endpoints of the SPS edge is similar. The distance threshold needs to be set according to the specific situation of the region of interest, and is generally set to a certain value. After setting the distance threshold, proceed according to the distance threshold. Search in sequence various points in the middle and PS edges that satisfy the distance threshold between PS points are added to the candidate PS edge set. In the middle. If the result is... If it is an empty set, then the point is moved from... The process involves removing PS candidate points and then proceeding to the "Are there still PS candidate points?" decision node to obtain a candidate PS edge set. Each edge in the candidate PS edge set undergoes phase difference analysis to eliminate atmospheric phase errors and improve image quality. A spectral estimation algorithm (such as Beamforming or Compressive Sensing) is then used for imaging. Next, a scattering point detection algorithm (such as Generalized Likelihood Ratio Test or Information Theoretic Criteria) is used to identify whether the two endpoints of a PS edge are SPS points. If both endpoints of a PS edge are SPS points and the similarity parameter between the two endpoints is not less than a preset similarity threshold, the PS edge is retained to construct a candidate SPS edge set; otherwise, the PS edge is deleted. An APS-similar edges are selected by setting a distance threshold, and the candidate edge set is then optimized to improve edge quality.

[0041] Step 106: Select the highest quality edge from the candidate SPS edge set and add it to the SPS edge set. Move the endpoint corresponding to the highest quality edge from the PS point set to the SPS point set to obtain the updated SPS edge set and SPS point set. Stop updating the SPS edge set and SPS point set when the PS point set is empty. Output the final SPS edge set and SPS point set. Construct the SAR tomographic reference network based on the final SPS edge set and SPS point set.

[0042] The highest quality edge is selected from the candidate SPS edge set and added to the SPS edge set, where the quality of the connecting edge is represented by the available signal ratio:

[0043]

[0044] in, This represents the differential signal between the endpoints of the connection. To estimate the obtained difference signal, It refers to the number of observations;

[0045] Simultaneously, the endpoints corresponding to the highest quality edges are moved from the PS point set to the SPS point set, resulting in the updated SPS edge set and SPS point set. Then, a check is performed to determine if there are still candidate PS points. If the value in the middle is not empty, continue repeating steps 104-106 until... If empty, the loop ends and outputs the final SPS edge set and SPS point set. New SPS points and SPS edges are continuously added to the SPS point set and SPS edge set. Each time an SPS point is added, only one SPS edge is added, and this edge has the highest imaging quality among all candidate SPS edges. The points and edges connected from the starting reference point based on the final SPS edge set and SPS point set are all optimal. Constructing a SAR tomographic reference network using optimal points and edges ensures maximum network coverage and optimal imaging quality. Figure 3 The image shown is an average SAR intensity map. The horizontal axis represents the azimuth direction, and the vertical axis represents the range direction. This map was obtained using measured data from TerraSAR-X, with the observation mode being 26 scenes in X-band Staring mode, and the azimuth and range resolutions being 0.24m and 0.59m, respectively. Figure 4 This is a reference network established using the method proposed in this application. The reference network generated based on the method proposed in this application can cover the entire scene, and all available PS points are connected. Figure 5 This is a height map of PS points identified based on the method proposed in this application. Since the reference net generated based on the method proposed in this application can cover the entire scene, the height of PS points in the entire scene can be effectively inverted.

[0046] In the aforementioned SAR tomographic reference network generation method based on minimum spanning tree, this application selects the starting reference point of the SAR tomographic reference network by calculating the amplitude deviation index value of all PS points in the imaging reference network constituting the remote sensing imaging area. Then, by setting PS edge search and SPS point judgment rules, it sequentially searches for PS edges in the SPS point set and PS point set that satisfy the distance threshold according to a pre-set distance threshold, and constructs a candidate PS edge set using the PS edges that satisfy the distance threshold. After performing phase difference on each edge in the candidate PS edge set, it performs imaging using a spectral estimation algorithm, and then detects the imaging results using a scattering point detection algorithm. The detection results are used to set a candidate SPS edge set, and finally, the candidate SPS edges are selected from the candidate SPS edges. The highest quality edges are selected and added to the SPS edge set. The endpoints corresponding to the highest quality edges are moved from the PS point set to the SPS point set, resulting in an updated SPS edge set and SPS point set. The update of the SPS edge set and SPS point set is stopped when the PS point set is empty. By continuously adding new SPS points and SPS edges to the SPS point set and SPS edge set, only one SPS edge is added each time an SPS point is added, and this edge has the highest imaging quality among all candidate SPS edges. Based on the final SPS edge set and SPS point set, the points and edges connected from the starting reference point are all optimal. Constructing a SAR tomographic reference network using the optimal points and edges can guarantee the maximum network coverage and the optimal imaging quality.

[0047] In one embodiment, the amplitude deviation index value of all PS points constituting the reference net is calculated, including:

[0048] The amplitude deviation index of all PS points constituting the reference net is calculated as follows:

[0049]

[0050] in, Indicates the number of observations. Represents a pixel The The amplitude value of the second observation.

[0051] In one embodiment, the preset distance threshold is set within a range of [range]. In one embodiment, based on a pre-set distance threshold, PS edges that satisfy the distance threshold are sequentially searched in the SPS point set and the PS point set, and a candidate PS edge set is constructed using the PS edges that satisfy the distance threshold, including:

[0052] Based on a pre-set distance threshold, sequentially search for PS edges in the SPS point set and PS point set that satisfy the distance threshold, and construct a candidate PS edge set using the PS edges that satisfy the distance threshold:

[0053]

[0054] in, For SPS point set The point in the middle, PS point set The point in the middle, , These are the spatial coordinates of the reference network vertex and the candidate vertex, respectively. This represents the Euclidean distance operator.

[0055] In one embodiment, the imaging results are detected according to a scattering point detection algorithm, and a candidate SPS edge set is set using the detection results, including:

[0056] The imaging results are detected using a scattering point detection algorithm. If both endpoints of the PS edge are SPS points and the similarity parameter of the two endpoints of the PS edge is not less than a preset similarity threshold, the PS edge is retained to construct a candidate SPS edge set; otherwise, the PS edge is deleted.

[0057] In one embodiment, the edge with the highest quality is:

[0058]

[0059] in, express Similarity parameters between them This represents the preset similarity threshold. This represents the endpoint corresponding to the edge with the highest quality.

[0060] In one embodiment, the endpoints corresponding to the highest quality edges are moved from the PS point set to the SPS point set, resulting in updated SPS edge sets and SPS point sets, including:

[0061] Move the endpoints corresponding to the highest quality edges from the PS point set to the SPS point set to obtain the updated SPS edge set and SPS point set as follows:

[0062]

[0063] in, .

[0064] It should be understood that, although Figure 1 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 1 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0065] In one embodiment, such as Figure 6 As shown, a SAR tomographic reference network generation device based on minimum spanning tree is provided, including: an initialization module 602, a candidate SPS edge set setting module 604, and a SAR tomographic reference network construction module 606, wherein:

[0066] Initialization module 602 is used to acquire the remote sensing imaging area, calculate the amplitude deviation index values ​​of all PS points in the imaging reference network constituting the remote sensing imaging area, select the PS point with the smallest amplitude deviation index value from all PS points as the starting reference point, and add the starting reference point to the SPS point set. The SPS edge set is initialized to an empty set. The imaging reference network includes the PS point set. An SPS point is a point that contains only one permanent scatterer in a single pixel.

[0067] The candidate SPS edge set setting module 604 is used to sequentially search for PS edges in the SPS point set and PS point set that meet the distance threshold according to a pre-set distance threshold, and to construct a candidate PS edge set using the PS edges that meet the distance threshold; after performing phase difference on each edge in the candidate PS edge set, imaging is performed using a spectral estimation algorithm, and then the imaging result is detected using a scattering point detection algorithm, and the candidate SPS edge set is set using the detection result.

[0068] The SAR tomographic reference network construction module 606 is used to select the highest quality edge from the candidate SPS edge set and add it to the SPS edge set, and move the endpoint corresponding to the highest quality edge from the PS point set to the SPS point set to obtain the updated SPS edge set and SPS point set. When the PS point set is empty, the update of the SPS edge set and SPS point set is stopped, and the final SPS edge set and SPS point set are output. The SAR tomographic reference network is constructed based on the final SPS edge set and SPS point set.

[0069] Specific limitations regarding the SAR tomographic reference network generation device based on minimum spanning tree can be found in the limitations of the SAR tomographic reference network generation method based on minimum spanning tree described above, and will not be repeated here. Each module in the aforementioned SAR tomographic reference network generation device based on minimum spanning tree can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0070] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0071] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for generating SAR tomographic reference networks based on minimum spanning trees, characterized in that, The method includes: The remote sensing imaging area is acquired, and the amplitude deviation index values ​​of all PS points in the imaging reference network constituting the remote sensing imaging area are calculated. Based on the amplitude deviation index values ​​of all PS points, the PS point with the smallest amplitude deviation index value is selected as the starting reference point, and the starting reference point is added to the SPS point set. The SPS edge set is initialized to an empty set. The imaging reference network includes the PS point set. The SPS point refers to a point that contains only one permanent scatterer in a single pixel. Based on a pre-set distance threshold, sequentially search for PS edges in the SPS point set and PS point set that satisfy the distance threshold, and construct a candidate PS edge set using the PS edges that satisfy the distance threshold; perform phase difference on each edge in the candidate PS edge set and then use a spectral estimation algorithm to perform imaging, and then use a scattering point detection algorithm to detect the imaging results, and use the detection results to set the candidate SPS edge set; Select the highest quality edge from the candidate SPS edge set and add it to the SPS edge set. Move the endpoint corresponding to the highest quality edge from the PS point set to the SPS point set to obtain the updated SPS edge set and SPS point set. Stop updating the SPS edge set and SPS point set when the PS point set is empty. Output the final SPS edge set and SPS point set. Construct the SAR tomographic reference network based on the final SPS edge set and SPS point set. The imaging results are detected using a scattering point detection algorithm, and the candidate SPS edge set is set based on the detection results, including: The imaging results are detected by the scattering point detection algorithm. If both endpoints of the PS edge are SPS points and the similarity parameter of the two endpoints of the PS edge is not less than the preset similarity threshold, the PS edge is retained to construct a candidate SPS edge set; otherwise, the PS edge is deleted. The edge with the highest quality is: in, express Similarity parameters between them This represents the preset similarity threshold. This represents the endpoint corresponding to the edge with the highest quality. For SPS point set The point in the middle, PS point set The point in the middle; Move the endpoints corresponding to the highest quality edges from the PS point set to the SPS point set to obtain the updated SPS edge set and SPS point set, including: Move the endpoints corresponding to the highest quality edges from the PS point set to the SPS point set to obtain the updated SPS edge set and SPS point set as follows: in, .

2. The method according to claim 1, characterized in that, Calculate the amplitude deviation index values ​​for all PS points constituting the reference network, including: The amplitude deviation index of all PS points constituting the reference net is calculated as follows: in, Indicates the number of observations. Represents a cell The The amplitude value of the second observation.

3. The method according to claim 1, characterized in that, The preset distance threshold is set within the following range: .

4. The method according to claim 3, characterized in that, Based on a pre-set distance threshold, sequentially search for PS edges in the SPS point set and PS point set that satisfy the distance threshold, and construct a candidate PS edge set using the PS edges that satisfy the distance threshold, including: Based on a pre-set distance threshold, sequentially search for PS edges in the SPS point set and PS point set that satisfy the distance threshold, and construct a candidate PS edge set using the PS edges that satisfy the distance threshold: in, For SPS point set The point in the middle, PS point set The point in the middle, , These are the spatial coordinates of the reference network vertex and the candidate vertex, respectively. This represents the Euclidean distance operator.

5. A SAR tomographic reference network generation device based on minimum spanning tree, characterized in that, The device includes: The initialization module is used to acquire the remote sensing imaging area, calculate the amplitude deviation index values ​​of all PS points in the imaging reference network constituting the remote sensing imaging area, select the PS point with the smallest amplitude deviation index value from all PS points as the starting reference point, and add the starting reference point to the SPS point set. The SPS edge set is initialized to an empty set. The imaging reference network includes the PS point set. The SPS point refers to a point that contains only one permanent scatterer in a single pixel. A candidate SPS edge set setting module is used to sequentially search for PS edges in the SPS point set and PS point set that satisfy the distance threshold according to a pre-set distance threshold, and construct a candidate PS edge set using the PS edges that satisfy the distance threshold; after performing phase difference on each edge in the candidate PS edge set, imaging is performed using a spectral estimation algorithm, and then the imaging result is detected using a scattering point detection algorithm. The candidate SPS edge set is set using the detection result, including: The imaging results are detected by the scattering point detection algorithm. If both endpoints of the PS edge are SPS points and the similarity parameter of the two endpoints of the PS edge is not less than the preset similarity threshold, the PS edge is retained to construct a candidate SPS edge set; otherwise, the PS edge is deleted. The SAR tomographic reference network construction module is used to select the highest quality edge from the candidate SPS edge set and add it to the SPS edge set, and move the endpoint corresponding to the highest quality edge from the PS point set to the SPS point set, obtaining the updated SPS edge set and SPS point set. The updating of the SPS edge set and SPS point set stops when the PS point set is empty, and the final SPS edge set and SPS point set are output. The SAR tomographic reference network is constructed based on the final SPS edge set and SPS point set. The highest quality edge is: in, express Similarity parameters between them This represents the preset similarity threshold. This represents the endpoint corresponding to the edge with the highest quality. For SPS point set The point in the middle, PS point set The point in the middle; Move the endpoints corresponding to the highest quality edges from the PS point set to the SPS point set to obtain the updated SPS edge set and SPS point set, including: Move the endpoints corresponding to the highest quality edges from the PS point set to the SPS point set to obtain the updated SPS edge set and SPS point set as follows: in, .

Images