Adaptive networking method, device and equipment

Abstract

The present application provides an adaptive network construction method, device and equipment, which relates to the field of geological monitoring technology. The method includes: triangulating multiple PS points on the interference diagram corresponding to the SAR image to obtain the corresponding initial triangulated network, wherein the initial triangulated network includes multiple PS points and multiple edges, and the two vertices of any edge are PS points; performing initial screening processing on the initial triangulated network to obtain a seed network, wherein the seed network is composed of seed PS points and seed edges, wherein the seed PS points are a subset of multiple PS points, and the seed edges are a subset of multiple edges; performing adaptive expansion and encryption processing on the seed network according to the seed PS points and seed edges to obtain an encrypted network; performing coherence screening processing and connectivity screening processing on the encrypted network to obtain a target network. The problem of losing measurement points (i.e., PS points) during the transmission unwrapping process due to the influence of phase incoherence caused by terrain undulation is overcome, which helps to improve the effect of subsequent PS-InSAR.

Description

Technical Field

[0001] The present application relates to the field of geological monitoring technology, and in particular to an adaptive networking method, device and equipment. Background Art

[0002] Surface deformation monitoring refers to monitoring the deformation degree of the surface through certain technical means. The deformation of the surface can include surface rise and surface fall, etc. Through surface deformation monitoring, the changes of the surface can be known, which is of great significance for geological disaster monitoring and other aspects.

[0003] At present, surface monitoring is usually carried out based on Persistent Scatterer Interferometric Synthetic Aperture Radar (PS-InSAR). PS-InSAR is a technology widely used in deformation monitoring and elevation correction in urban areas, landslides, etc. It monitors surface deformation by analyzing the time-stable permanent scatterers (PS) in the Synthetic Aperture Radar (SAR) image sequence. It can achieve low-cost, high-precision and large-scale surface monitoring, which has great advantages.

[0004] In surface monitoring through PS-InSAR, effective network construction is the prerequisite for achieving surface monitoring. Network construction is the process of building a network model on the interference pattern composed of PS points. The current PS-InSAR technology is easily affected by the terrain in mountainous areas, which may lead to the loss of PS points during the transmission unwrapping process. Summary of the invention

[0005] The present application provides an adaptive network construction method, device and equipment to solve the problem that the current PS-InSAR technology is easily affected by the terrain undulation in mountainous areas and causes the loss of PS points during the transmission unwrapping process.

[0006] In a first aspect, the present application provides an adaptive networking method, comprising:

[0007] Performing triangulation on a plurality of permanent scatterer PS points on an interference pattern corresponding to a synthetic aperture radar SAR image to obtain a corresponding initial triangulation network, wherein the initial triangulation network includes the plurality of PS points and a plurality of edges, and two vertices of any edge are both the PS points;

[0008] Performing initial screening processing on the initial triangulated network to obtain a seed network, wherein the seed network is composed of seed PS points and seed edges, wherein the seed PS points are a subset of the multiple PS points, and the seed edges are a subset of the multiple edges;

[0009] Performing adaptive extended encryption processing on the seed network according to the seed PS point and the seed edge to obtain an encrypted network;

[0010] The encrypted network is subjected to coherence screening processing and connectivity screening processing to obtain a target network.

[0011] In a possible implementation manner, triangulating a plurality of permanent scatterer PS points on an interference pattern corresponding to a synthetic aperture radar SAR image to obtain a corresponding initial triangulated network includes:

[0012] determining positions of the plurality of PS points on the interference pattern;

[0013] According to the positions of the multiple PS points on the interference graph, Delaunay triangulation is performed on the multiple PS points to obtain the initial triangulation network;

[0014] Among them, the initial triangulated network includes multiple triangles, and any triangle is composed of three PS points and edges connecting the three PS points in pairs; the circumscribed circle of any triangle does not contain any PS points other than the three PS points that constitute the triangle; after the diagonals of the convex quadrilateral formed by any two adjacent triangles are exchanged with each other, the minimum angle of the six internal angles of the convex quadrilateral no longer increases.

[0015] In a possible implementation, the initial screening process is performed on the initial triangulated network to obtain a seed network, including:

[0016] Executing a first operation, the first operation comprising: updating the PS points in the triangulated network updated for the i-1th time to obtain the triangulated network updated for the ith time;

[0017] Performing a second operation, the second operation comprising: determining the number of PS points in the triangulated network after the i-th update;

[0018] When the number of PS points in the triangulated network after the i-th update is less than the number of PS points in the triangulated network after the i-1-th update, update i to i+1, and repeat the first operation and the second operation; wherein, i is initially 1, i is a positive integer greater than or equal to 1, and the triangulated network after the 0th update is the initial triangulated network;

[0019] When the number of PS points in the triangulated mesh after the i-th update is equal to the number of PS points in the triangulated mesh after the (i-1)-th update, the triangulated mesh after the i-th update is determined as the seed mesh.

[0020] In a possible implementation manner, the updating of the PS points in the triangulated network updated for the i-1th time to obtain the triangulated network updated for the ith time includes:

[0021] Obtaining the time coherence of each edge in the triangulated network after the (i-1)th update;

[0022] According to the temporal coherence and the first coherence threshold, removing low-quality edges in the triangulated network after the (i-1)th update, the low-quality edges being edges whose corresponding temporal coherence is less than or equal to the first coherence threshold;

[0023] According to the low-quality edge, isolated PS points in the triangulated network after the (i-1)th update are removed to obtain non-isolated PS points in the triangulated network after the (i-1)th update, wherein the isolated PS points are PS points that are not connected to any edge other than the low-quality edge;

[0024] Delaunay triangulation is performed according to the non-isolated PS points to obtain the triangulated network after the i-th update.

[0025] In a possible implementation manner, the step of performing adaptive extended encryption processing on the seed network according to the seed PS point and the seed edge to obtain an encrypted network includes:

[0026] Performing statistical processing on the seed edges associated with each of the seed PS points to obtain the number of seed edges associated with each of the seed PS points in different quadrants;

[0027] Determine the quadrant to be encrypted of each seed PS point according to the number of seed edges associated with each seed PS point in different quadrants;

[0028] The seed network is adaptively extended and encrypted according to the seed PS points in the to-be-encrypted quadrant of each seed PS point to obtain the encrypted network.

[0029] In a possible implementation manner, performing adaptive extended encryption processing on the seed network according to the seed PS points in the to-be-encrypted quadrant of each seed PS point to obtain the encrypted network includes:

[0030] For any seed PS point x, determine the PS points to be connected to the seed PS point x according to the distances between the seed PS point x and other PS points in the quadrant to be encrypted of the seed PS point x;

[0031] Determine the edge to be connected according to the seed PS point x and the PS point to be connected;

[0032] The seed network is adaptively extended and encrypted according to the to-be-connected edges corresponding to each of the seed PS points to obtain the encrypted network.

[0033] In a possible implementation manner, performing coherence screening processing and connectivity screening processing on the encrypted network to obtain a target network includes:

[0034] Determining a first type of PS point and a second type of PS point in the encrypted network according to a second coherence threshold and a first connectivity threshold;

[0035] Acquire a skeleton network corresponding to the encrypted network according to the first type of PS points;

[0036] According to a third coherence threshold and a second connectivity threshold, the second type of PS points are connected to the skeleton network to obtain the target network, wherein the third coherence threshold is less than or equal to the second coherence threshold, and the second connectivity threshold is less than or equal to the first connectivity threshold.

[0037] In a second aspect, an embodiment of the present application provides an adaptive networking device, including:

[0038] A meshing module is used to triangulate the multiple PS points on the interference pattern corresponding to the SAR image to obtain a corresponding initial triangulated network, wherein the initial triangulated network includes the multiple PS points and multiple edges, and the two vertices of any edge are the PS points;

[0039] A processing module, used for performing initial screening processing on the initial triangulated network to obtain a seed network, wherein the seed network is composed of seed PS points and seed edges, wherein the seed PS points are a subset of the plurality of PS points, and the seed edges are a subset of the plurality of edges;

[0040] An encryption module, used for performing adaptive extended encryption processing on the seed network according to the seed PS point and the seed edge to obtain an encrypted network;

[0041] The screening module is used to perform coherence screening and connectivity screening on the encrypted network to obtain a target network.

[0042] In a possible implementation manner, the networking module is specifically used to:

[0043] determining positions of the plurality of PS points on the interference pattern;

[0044] According to the positions of the multiple PS points on the interference graph, Delaunay triangulation is performed on the multiple PS points to obtain the initial triangulation network;

[0045] Among them, the initial triangulated network includes multiple triangles, and any triangle is composed of three PS points and edges connecting the three PS points in pairs; the circumscribed circle of any triangle does not contain any PS points other than the three PS points that constitute the triangle; after the diagonals of the convex quadrilateral formed by any two adjacent triangles are exchanged with each other, the minimum angle of the six internal angles of the convex quadrilateral no longer increases.

[0046] In a possible implementation manner, the processing module is specifically used to:

[0047] Executing a first operation, the first operation comprising: updating the PS points in the triangulated network updated for the i-1th time to obtain the triangulated network updated for the ith time;

[0048] Performing a second operation, the second operation comprising: determining the number of PS points in the triangulated network after the i-th update;

[0049] When the number of PS points in the triangulated network after the i-th update is less than the number of PS points in the triangulated network after the i-1-th update, update i to i+1, and repeat the first operation and the second operation; wherein, i is initially 1, i is a positive integer greater than or equal to 1, and the triangulated network after the 0th update is the initial triangulated network;

[0050] When the number of PS points in the triangulated mesh after the i-th update is equal to the number of PS points in the triangulated mesh after the (i-1)-th update, the triangulated mesh after the i-th update is determined as the seed mesh.

[0051] In a possible implementation manner, the processing module is specifically used to:

[0052] Obtaining the time coherence of each edge in the triangulated network after the (i-1)th update;

[0053] According to the temporal coherence and the first coherence threshold, removing low-quality edges in the triangulated network after the (i-1)th update, the low-quality edges being edges whose corresponding temporal coherence is less than or equal to the first coherence threshold;

[0054] According to the low-quality edge, isolated PS points in the triangulated network after the (i-1)th update are removed to obtain non-isolated PS points in the triangulated network after the (i-1)th update, wherein the isolated PS points are PS points that are not connected to any edge other than the low-quality edge;

[0055] Delaunay triangulation is performed according to the non-isolated PS points to obtain the triangulated network after the i-th update.

[0056] In a possible implementation manner, the encryption module is specifically used for:

[0057] Performing statistical processing on the seed edges associated with each of the seed PS points to obtain the number of seed edges associated with each of the seed PS points in different quadrants;

[0058] Determine the quadrant to be encrypted of each seed PS point according to the number of seed edges associated with each seed PS point in different quadrants;

[0059] The seed network is adaptively extended and encrypted according to the seed PS points in the to-be-encrypted quadrant of each seed PS point to obtain the encrypted network.

[0060] In a possible implementation manner, the encryption module is specifically used for:

[0061] For any seed PS point x, determine the PS points to be connected to the seed PS point x according to the distances between the seed PS point x and other PS points in the quadrant to be encrypted of the seed PS point x;

[0062] Determine the edge to be connected according to the seed PS point x and the PS point to be connected;

[0063] The seed network is adaptively extended and encrypted according to the to-be-connected edges corresponding to each of the seed PS points to obtain the encrypted network.

[0064] In a possible implementation, the screening module is specifically used for:

[0065] Determining a first type of PS point and a second type of PS point in the encrypted network according to a second coherence threshold and a first connectivity threshold;

[0066] Acquire a skeleton network corresponding to the encrypted network according to the first type of PS points;

[0067] According to a third coherence threshold and a second connectivity threshold, the second type of PS points are connected to the skeleton network to obtain the target network, wherein the third coherence threshold is less than or equal to the second coherence threshold, and the second connectivity threshold is less than or equal to the first connectivity threshold.

[0068] In a third aspect, the present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the adaptive networking method as described in any one of the first aspects is implemented.

[0069] In a fourth aspect, the present application provides a non-transitory computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the adaptive networking method as described in any one of the first aspects is implemented.

[0070] In a fifth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the adaptive networking method as described in any one of the first aspects.

[0071] The adaptive meshing method, device and equipment provided by the present application first perform triangulation on multiple PS points on the interference pattern corresponding to the SAR image to obtain a corresponding initial triangulation network, wherein the initial triangulation network includes multiple PS points and multiple edges, and the two vertices of any edge are PS points; then, the initial triangulation network is initially screened to obtain a seed network, which is composed of seed PS points and seed edges, wherein the seed PS points are a subset of multiple PS points, and the seed edges are a subset of multiple edges; the seed network is adaptively extended and encrypted according to the seed PS points and seed edges to obtain an encrypted network, and the encrypted network is screened for coherence and connectivity to obtain a target network. After the initial triangulated network is obtained through triangulation, the initial triangulated network is initially screened to achieve preliminary optimization of the initial triangulated network. Then, the degree of network imbalance in the region is reduced through adaptive expansion and encryption processing. Finally, the PS points and edges with low coherence and connectivity in the encrypted network can be removed through coherence screening and connectivity screening processing. This method is suitable for mountainous areas where the deformation rate cannot be calculated due to terrain undulations. It overcomes the problem of losing measurement points (i.e., PS points) during the transfer unwrapping process due to the influence of phase incoherence caused by terrain undulations, and helps to improve the effect of subsequent PS-InSAR. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] In order to more clearly illustrate the technical solutions in the present application or the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0073] Figure 1 A schematic diagram of a process flow of an adaptive networking method provided in an embodiment of the present application;

[0074] Figure 2 A schematic diagram of a triangulated network provided in an embodiment of the present application;

[0075] Figure 3 The seed PS point and seed edge network length result diagram provided in the embodiment of the present application;

[0076] Figure 4 A coherence result diagram of the seed PS point and seed edge network provided in the embodiment of the present application;

[0077] Figure 5 The adaptive extended encryption edge length result graph provided in the embodiment of the present application;

[0078] Figure 6 A schematic diagram of proximity networking and distance threshold networking provided in an embodiment of the present application;

[0079] Figure 7 The optimized skeleton network side length result diagram provided in the embodiment of the present application;

[0080] Figure 8 The result diagram of the optimized skeleton network coherence provided in the embodiment of the present application;

[0081] Fig. 9 A schematic diagram of the structure of an adaptive networking device provided in an embodiment of the present application;

[0082] Fig.10 A schematic diagram of the physical structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0083] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0084] PS-InSAR refers to the technology of monitoring surface deformation by analyzing the time-stable permanent scatterers (PS) in the SAR image sequence. It is widely used in deformation monitoring and elevation correction in urban areas, landslides, etc. The key idea of ​​PS-InSAR technology is to use multiple SAR images (SAR image time series) acquired in the same area within a certain period of time, and use statistical analysis methods to detect targets with high temporal correlation (i.e. PS) in the imaging area, and then model and analyze the phase time series of these PS targets to separate deformation and atmospheric delay information. This modeling and processing strategy is very beneficial to reduce or overcome the influence of temporal and spatial decorrelation and atmospheric delay.

[0085] PS-InSAR technology overcomes the shortcomings of traditional single-point deformation measurement technology and synthetic aperture radar differential interferometry technology, and can achieve low-cost, high-precision and large-scale surface monitoring. It has great advantages and is better used in urban areas with many strong scattering targets. However, in non-urban areas, bare land, and vegetation-covered areas, the measurement points are lost during the transmission unwrapping process due to the phase decoherence of terrain undulation, resulting in the failure of unwrapping and the inability to invert the deformation rate of the measurement point.

[0086] Optimizing the network model is a crucial step to improve PS-InSAR technology; the idea of ​​network construction is to form a stable reference network between PS points. Due to the spatial proximity between adjacent measurement points, the atmospheric and orbital environment differences have the same atmospheric and orbital environment, which can better weaken the atmospheric and orbital effects, and there is no need to compensate for the initial processing of the spatially related phase. Therefore, this feature can be used to estimate the linear deformation rate and elevation error of each PS pixel in the network using a linear model, and then the nonlinear component of the motion and the atmospheric phase of each image are calculated using spatiotemporal filtering. Based on the redundant network, the reliability and accuracy of the unwrapping results can be ensured. This method can solve the deformation information of more PS points in the network, and the deformation inversion accuracy is higher.

[0087] The main development process of the meshing process can be divided into three stages: (1) Single reference meshing, which is to differentiate all pixels from the reference point. This method is only applicable to pixels within a few hundred meters of the reference point. When the distance between the PS pixel and the reference pixel is far, the error exceeds the limit and more PS points cannot be solved. (2) Transfer meshing, which is a way to combine local reference points and global reference points. The relative difference between the calculated points is expanded in the form of region growth. The errors generated in the transfer process are also accumulated, resulting in unreliable results in the later untangling process. (3) Redundant meshing. A highly redundant meshing graph is obtained from the connected point pairs, and a solution that is more robust to noise and outliers is obtained. The development process of redundant networks includes the following aspects: 1) Delaunay triangulation, the simplest and most commonly used meshing method. This meshing method seeks whether the convex hull of two adjacent triangles contains other points and whether the triangles overlap. Therefore, it is independent of the phase continuity rule. This means that when there are deformation signals, atmospheric turbulence and phase noise in the interference pattern, many edges in the spatial network may have high phase differences and cannot be correctly untangled. 2) Set a distance threshold to connect with surrounding points, such as the nearest neighbor algorithm (K-nearest neighbors, KNN), connect surrounding points according to a specific distance threshold to form a redundant network. The disadvantage of this method is that if the distance threshold is small, points far away cannot be connected. If the distance threshold is too large, the network may be too redundant, resulting in low computational time efficiency. 3) Consider temporal coherence to build a network. Due to the limitations of the Delaunay triangulation, loss of coherence causes the phase difference to be greater than π and cannot be untangled. Therefore, temporal coherence is used to quantify the phase difference. The higher the temporal coherence, the smaller the phase change of an edge in the time dimension. Selecting edges with higher coherence is conducive to improving the phase untangling results of the interference graph, but if non-high coherence edges are removed, the connected network will be interrupted. 4) Consider the connectivity threshold to build a network. The disadvantage of the distance threshold network is that some points may not be connected. Setting the connectivity threshold ensures that the target point and surrounding points are all connected. However, since each point is connected with the same connectivity threshold, the part with more concentrated points causes edge redundancy, which increases the amount of calculation.

[0088] As for the uneven distribution of points in the study area, it can be seen that whether using Delaunay triangulation alone or combining distance threshold, coherence, and connectivity threshold to construct a network, it is impossible to balance the number of edges in areas with dense points and areas with sparse points. Generally speaking, increasing the connectivity threshold will inevitably lead to serious edge redundancy in areas with dense points, while reducing the connectivity threshold will cause areas with sparse points to be unable to connect and construct a network.

[0089] Based on this, the embodiment of the present application provides an adaptive network construction method. In view of the current triangulation network that only considers the network construction characteristics but ignores the shortcoming that the height difference between the actual two points is too large to be disentangled, the triangulation network is optimized by using the distance threshold and time coherence. At the same time, in view of the uneven distribution of points in the study area, the network is adaptively encrypted using the connectivity threshold in areas with sparse edges, while the network encryption operation is not performed using the connectivity threshold in areas with dense edges. In addition, considering the connectivity, coherence and number of edges, some edges are adaptively added to optimize the network, and a redundant network is adaptively constructed. The following is combined with Figure 1 The solution of the embodiment of the present application is introduced.

[0090] It should be noted that the execution subject of each embodiment of the present application can be a processor, a microprocessor, a server, a chip, etc., or it can be a device integrating the above-mentioned processor, microprocessor or chip and other devices, such as a client or terminal device. The specific execution subject of each embodiment of the present application is not limited, and it can be selected and set according to actual needs.

[0091] Figure 1 A flow chart of the adaptive networking method provided in the embodiment of the present application is shown in FIG. Figure 1 As shown, the method may include:

[0092] S11, triangulating a plurality of PS points on the interference pattern corresponding to the SAR image to obtain a corresponding initial triangulated network, wherein the initial triangulated network includes a plurality of PS points and a plurality of edges, and two vertices of any edge are both PS points.

[0093] SAR images are images produced by the SAR system. The SAR system is an active earth observation system that can be installed on aircraft, satellites, spacecraft and other flight platforms to observe the earth around the clock and in all weather conditions. The SAR system uses the principle of synthetic aperture to achieve high-resolution microwave imaging. The SAR system has a certain surface penetration capability and usually only records the echo information of one band. The complex data of each pixel on the SAR image can be transformed to extract the corresponding amplitude and phase information.

[0094] PS points refer to targets with stable scattering characteristics in the imaging area. Their radar backscattering characteristics are not easy to change with time. For example, hard targets such as houses, bridges, railway ballasts, concrete dams, rocks, etc. can all be identified as PS points. PS points are all strong scattering point targets and are less affected by time and baseline decorrelation, thereby maintaining a high signal-to-noise ratio for phase observations.

[0095] PS-InSAR technology refers to the technology of interferometric measurement using the phase information in SAR images of different periods in the same area. Based on PS-InSAR technology, multiple SAR images are firstly subjected to registration, radiation calibration, PS detection and interferometric processing to obtain the interference map corresponding to the SAR image.

[0096] The interference graph includes multiple PS points. The multiple PS points on the interference graph are triangulated to obtain the corresponding initial triangulated network. Triangulation is the process of connecting multiple PS points to form multiple triangles. The initial triangulated network includes multiple triangles, and the three vertices of each triangle are PS points, that is, the initial triangulated network includes multiple PS points and multiple edges, and the two vertices of any edge are PS points.

[0097] S12, performing initial screening processing on the initial triangulated network to obtain a seed network, wherein the seed network is composed of seed PS points and seed edges, wherein the seed PS points are a subset of multiple PS points, and the seed edges are a subset of multiple edges.

[0098] After obtaining the initial triangulated network, the initial triangulated network can be subjected to initial screening processing. The purpose of the initial screening processing is to screen out high-quality PS points and edges in the initial triangulated network. These high-quality PS points are seed PS points, and high-quality edges are seed edges. Seed PS points are a subset of multiple PS points in the initial triangulated network, and seed edges are a subset of multiple edges in the initial triangulated network.

[0099] The initial screening process can be implemented based on the coherence of the edges in the initial triangulated network. For example, the coherence of each edge in the initial triangulated network can be calculated, and then a certain coherence threshold can be set. The edges in the initial triangulated network can be screened according to the coherence threshold to remove low-quality edges. After removing the low-quality edges, there may be some isolated PS points, that is, PS points that are not connected to any edges. These PS points are low-quality PS points. Then remove such PS points. The remaining ones are high-quality PS points and high-quality edges, that is, seed PS points and seed edges. The seed PS points and seed edges constitute a seed network.

[0100] S13, performing adaptive expansion and encryption processing on the seed network according to the seed PS points and the seed edges to obtain an encrypted network.

[0101] From the initial triangulated network to the seed network, some low-quality PS points and edges are removed, so some PS points are not included in the seed network. At the same time, there is an imbalance that some areas have dense edges and other areas have sparse edges. Therefore, after obtaining the seed network, the seed network can be adaptively expanded and encrypted according to the seed PS points and seed edges.

[0102] For example, the distribution of the edges connecting each seed PS point in the seed network with other seed PS points can be counted. If there are fewer edges in some areas, they can be adaptively connected to the seed PS points in the area to increase the edges in the area. If there are more edges in some areas, they can be left unconnected. Through the above encryption processing, the adaptability of the encrypted network can be enhanced.

[0103] S14, performing coherence screening and connectivity screening on the encrypted network to obtain a target network.

[0104] After obtaining the encrypted network, the encrypted network can be subjected to coherence screening and connectivity screening processing to remove edges with low coherence and PS points with low connectivity, so as to further optimize the encrypted network and finally obtain the target network, which can be used for subsequent PS-InSAR processing.

[0105] The adaptive meshing method provided in the embodiment of the present application first performs triangulation on multiple PS points on the interference pattern corresponding to the SAR image to obtain a corresponding initial triangulation network, wherein the initial triangulation network includes multiple PS points and multiple edges, and the two vertices of any edge are PS points; then, the initial triangulation network is initially screened to obtain a seed network, wherein the seed network is composed of seed PS points and seed edges, wherein the seed PS points are a subset of multiple PS points, and the seed edges are a subset of multiple edges; the seed network is adaptively extended and encrypted according to the seed PS points and the seed edges to obtain an encrypted network, and the encrypted network is screened for coherence and connectivity to obtain a target network. After the initial triangulated network is obtained through triangulation, the initial triangulated network is initially screened to achieve preliminary optimization of the initial triangulated network. Then, the degree of network imbalance in the region is reduced through adaptive expansion and encryption processing. Finally, the PS points and edges with low coherence and connectivity in the encrypted network can be removed through coherence screening and connectivity screening processing. This method is suitable for mountainous areas where the deformation rate cannot be calculated due to terrain undulations. It overcomes the problem of losing measurement points (i.e., PS points) during the transfer unwrapping process due to the influence of phase incoherence caused by terrain undulations, and helps to improve the effect of subsequent PS-InSAR.

[0106] Based on the introduction of the above embodiments, the solution of the present application is further introduced below in conjunction with the accompanying drawings and the PS-InSAR technical process.

[0107] Atmospheric delay is manifested as spatial autocorrelation of hundreds of meters in spatial scale. Therefore, although the atmospheric state of different real-phase SAR images is inconsistent, the atmospheric state of adjacent target points in the same image area shows a high similarity. The closer the ground points are, the higher the correlation of atmospheric state. After extracting the coherence points from the time-series SAR images, the candidate points are connected according to the optimized network, and differential phase connection is established for each target point that meets the distance conditions to eliminate the influence of the atmosphere.

[0108] The solved linear model is adjusted to the phase increment between two adjacent pixels connected to the optimization network, avoiding the need for sparse grid phase unwrapping of the interferogram. The SAR signal of the candidate point is compared with the signal of its neighboring points, and no preprocessing of the data amplitude and phase calibration is required. The atmospheric, orbital and spatial perturbations between adjacent candidate points are roughly the same.

[0109] A series of registered SAR images, the edge a formed by connecting two adjacent points in the interference graph between the i-th image and the main image, the phase difference between these two points is defined as δφ a,i ,δh a and δv a Respectively represent the residual elevation difference and velocity difference between two points, B i and T i Respectively represent the vertical baseline and the time baseline between the i-th image and the main image, δφ a,i It can be expressed by the following formula (1):

[0110]

[0111] λ represents the wavelength, ε a,i It represents the thermal noise and speckle noise between two points, which can be ignored when the distance between the two points is very close.

[0112] Formula (1) The phase difference between the two points connected by edge a can be expressed as the equation of residual elevation difference and velocity difference. Here, a parameter of coherence related to edge a is introduced to test the edge, which is defined as:

[0113]

[0114] w a,i represents the weight value calculated according to the amplitude value, coherence γ a Used to calculate the quality of the edge between points. Generally speaking, if γ a If it is greater than a given threshold, the edge should be retained.

[0115] The first step of the PS-InSAR technical process is to preprocess the SAR image.

[0116] The preprocessing of SAR images includes registration, circling the area of ​​interest and performing cropping operations. Since SAR images are usually for a very large area, and terrain monitoring is usually a small area, cropping operations are required to select the area of ​​interest. At the same time, some basic interferometric processing steps are performed, such as multi-viewing, generating intensity maps, generating digital elevation models (DEM) maps and geocoding. In addition, the preprocessing process depends on the spatiotemporal baseline of the collected SAR image data to select the appropriate main image.

[0117] The second step of the PS-InSAR technical process is point selection, that is, selecting PS points.

[0118] PS points are pixels with stable phases on the image. The stability of the phase can be represented by the stability of the amplitude on the SAR image. These points are usually part of man-made objects, such as buildings and metal structures, or other natural objects, such as exposed rocks.

[0119] The third step of the PS-InSAR technical process is to build a network model.

[0120] For the interference graph composed of PS point targets, the digital elevation model (DEM) is first used to remove the flat ground and terrain phases, and then the PS points are networked and connected according to a certain rule to construct a PS point neighborhood differential network model so that the interference of phase atmospheric delay and spatial decorrelation on the phase is more effectively suppressed. The solution of the embodiment of the present application is mainly aimed at the improvement of this step.

[0121] The key to building a high-quality adaptive network is based on the length of the edge, the number of edges, the coherence between edges, network connectivity, and spatial adaptability. There are five main criteria for selecting edges: 1) The edge length cannot be too long to avoid atmospheric and other space-related disturbances; 2) The number of edges. The number of edges cannot be too many to avoid the increase in computational complexity and affect computational efficiency; 3) Edge coherence. Edge coherence ensures the reliability of edge estimation. Generally speaking, the edges between adjacent candidate points have high coherence; 4) Network connectivity is the key to ensure that the networks in the study area can be smoothly disentangled; 5) Spatial adaptability. Balance the number of edges in space, increase the number of edges in edge-sparse areas, and reduce the number of edges in edge-dense areas.

[0122] After obtaining the interference graph corresponding to the SAR image, the positions of multiple PS points on the interference graph are first determined, and then Delaunay triangulation is performed on the multiple PS points according to the positions of the multiple PS points on the interference graph to obtain the initial triangulation network.

[0123] The Delaunay triangulation principles include: the circumscribed circle of each Delaunay triangle does not contain any other points in the surface; the diagonals of the convex quadrilateral formed by every two adjacent triangles will no longer increase the minimum angle of the six internal angles after being exchanged. Figure 2 A schematic diagram of a triangular network provided in an embodiment of the present application is shown in FIG. Figure 2 As shown in the figure, after Delaunay triangulation, an initial triangulation can be obtained. The initial triangulation includes multiple triangles. Any triangle is composed of three PS points and edges connecting the three PS points in pairs. The circumscribed circle of any triangle does not contain other PS points except the three PS points that constitute the triangle. After the diagonals of the convex quadrilateral formed by any two adjacent triangles are exchanged with each other, the minimum angle of the six internal angles of the convex quadrilateral no longer increases.

[0124] After obtaining the initial triangulated network, the initial triangulated network can be subjected to initial screening processing to obtain a seed network. The following is an introduction to this process.

[0125] Execute a first operation, the first operation comprising: updating the PS points in the triangulated network updated for the i-1th time to obtain the triangulated network updated for the ith time.

[0126] Specifically, first obtain the temporal coherence of each edge in the triangulated network after the i-1th update, and then remove low-quality edges in the triangulated network after the i-1th update based on the temporal coherence and the first coherence threshold, where the low-quality edges are edges whose corresponding temporal coherence is less than or equal to the first coherence threshold.

[0127] Furthermore, isolated PS points in the triangulated network after the i-1th update can be removed according to the low-quality edges to obtain non-isolated PS points in the triangulated network after the i-1th update. Isolated PS points are PS points that are not connected to edges other than low-quality edges.

[0128] The Delaunay triangulation is performed based on non-isolated PS points to obtain the i-th updated triangulated network.

[0129] After obtaining the triangulated network after the i-th update, a second operation is performed, and the second operation includes: determining the number of PS points in the triangulated network after the i-th update.

[0130] When the number of PS points in the triangulated network after the i-th update is less than the number of PS points in the triangulated network after the i-1-th update, update i to i+1, and repeat the first operation and the second operation; wherein i is initially 1, i is a positive integer greater than or equal to 1, and the triangulated network after the 0th update is the initial triangulated network;

[0131] When the number of PS points in the triangulated mesh after the i-th update is equal to the number of PS points in the triangulated mesh after the (i-1)-th update, the triangulated mesh after the i-th update is determined as a seed mesh.

[0132] As mentioned above, the process of obtaining the seed network based on the initial triangulated network is a cyclic operation process. Specifically, the temporal coherence of each edge in the initial triangulated network is first calculated, and then according to the first coherence threshold, the edges whose temporal coherence is less than or equal to the first coherence threshold are masked, and these edges are low-quality edges. On the basis of masking the low-quality edges, the isolated PS points can be temporarily masked, and the remaining PS points are retained. The remaining PS points are the non-isolated PS points in this round of update, and the isolated PS points are the PS points that are not connected to any edge after masking the low-quality edges.

[0133] Then, the remaining PS points are subjected to Delaunay triangulation to obtain the first updated triangulation. If the number of PS points in the first updated triangulation is less than the number of PS points in the initial triangulation, the above operation is repeated for the first updated triangulation, that is, the coherence of the edges in the first updated triangulation is calculated, low-quality edges are removed according to the first coherence threshold, and isolated PS points are removed based on the low-quality edges, so as to obtain the second updated triangulation by performing Delaunay triangulation based on the remaining PS points.

[0134] Similarly, if the number of PS points in the triangulated mesh after the second update is less than the number of PS points in the triangulated mesh after the first update, repeat the above operation on the triangulated mesh after the second update; if the number of PS points in the triangulated mesh after the second update is equal to the number of PS points in the triangulated mesh after the first update, determine the triangulated mesh after the second update as the seed mesh.

[0135] That is, when the number of PS points in the currently updated triangulated network is less than the number of PS points in the last updated triangulated network, it is necessary to re-execute the above operation on the currently updated triangulated network to obtain the next updated triangulated network; when the number of PS points in the currently updated triangulated network is equal to the number of PS points in the last updated triangulated network, the currently updated triangulated network is determined as the seed network.

[0136] Figure 3 The seed PS point and seed edge network length result graph provided in the embodiment of the present application includes both edges with shorter lengths and edges with longer lengths. Figure 4 The seed PS point and seed edge network coherence result diagram provided in the embodiment of the present application is as follows: Figure 4 As shown, some of the seed edges have lower temporal coherence, while some of the seed edges have higher temporal coherence.

[0137] After obtaining the seed network, the seed network can be adaptively extended and encrypted to obtain an encrypted network. For any seed PS point x, the PS points to be connected to the seed PS point x can be determined based on the distances between other PS points in the quadrant to be encrypted of the seed PS point x and the seed PS point x. Then, the edges to be connected are determined based on the seed PS point x and the PS points to be connected, so that the seed network is adaptively extended and encrypted based on the edges to be connected corresponding to various sub-PS points to obtain an encrypted network.

[0138] Specifically, in the process of repeatedly constructing the Delaunay triangulation network to obtain the seed network, the edge set associated with each seed PS point can be cumulatively recorded as an initial sampling of the edge set to be selected, and then the edges associated with each seed PS point x are divided into different quadrants according to their directions (for example, 8 quadrants can be divided), and the seed edges distributed in different quadrants are statistically analyzed to calculate the maximum edge length. Finally, according to the construction rules of the Delaunay triangulation network, it can be seen that if the edge length value is longer, it means that the seed PS point x has fewer adjacent PS points in this direction, otherwise, there are more adjacent PS points.

[0139] Based on the above statistical analysis results, for each seed PS point x, the edge length R^k_t=min(alpha*R^k_max,R_c) is used as the threshold in quadrant k to connect other PS points to achieve an adaptive encrypted network, where alpha is an adjustable scaling factor and R_c is the upper limit of the edge length, which is generally set to the correlation length of the atmospheric phase, for example, it can be set to about 1km. In actual processing, if R^i_max=0, then R^i_t=R_c.

[0140] Repeat the above process for all seed PS points to obtain all the edges to be connected. Finally, calculate the temporal coherence of the edges to be connected, remove the longer and low-quality edges to be connected, and remove the isolated PS points to obtain the encrypted network. Figure 5 The adaptive extended encryption edge length result diagram provided in the embodiment of the present application is as follows: Figure 6 The schematic diagram of the proximity network and distance threshold network provided in the embodiment of the present application is as follows: Figure 5 and Figure 6 As shown, the adaptive extended encryption networking method provided in the embodiment of the present application is compared with Figure 6 The example proximity network and distance threshold network construction methods show uniform distribution, short edge length, less redundancy and obvious adaptability.

[0141] After obtaining the encrypted network, the first type of PS points and the second type of PS points are determined in the encrypted network according to the second coherence threshold and the first connectivity threshold, wherein the edges in the encrypted network can be screened according to the second coherence threshold and the first connectivity threshold, and the edges whose temporal coherence is less than or equal to the second coherence threshold and whose connectivity is less than or equal to the first connectivity threshold are removed. After removing these edges, the isolated PS points are determined as the second type of PS points, and the other PS points are the first type of PS points.

[0142] Then, a skeleton network corresponding to the encrypted network is obtained according to the first type of PS points, wherein the skeleton network is adjusted on the basis of the encrypted network, for example, the skeleton network can be obtained by triangulating the network according to the first type of PS points.

[0143] Then, according to the third coherence threshold and the second connectivity threshold, the second type of PS points are connected to the skeleton network to obtain the target network, which is the final maximum connected subset network. Specifically, all the edges obtained in the above processes can be stored in a table. Then, with the PS points not included in the skeleton network as the center, some edges with slightly lower quality are found from the table, but the edges still need to satisfy their temporal coherence greater than or equal to the third coherence threshold, and the edges must be connected to the skeleton points (skeleton points are PS points on the skeleton network), and it is necessary to ensure that the number of connected skeleton points is sufficient (that is, the number of connected skeleton points is greater than or equal to the second connectivity threshold).

[0144] The third coherence threshold is less than or equal to the second coherence threshold, and the second connectivity threshold is less than or equal to the first connectivity threshold. The second coherence threshold and the first connectivity threshold can be appropriately adjusted according to the actual research area.

[0145] Figure 7 The optimized skeleton network side length result diagram provided in the embodiment of the present application shows that the side length of the network connection in the diagram is relatively short. Figure 8 The optimized skeleton network coherence result diagram provided in the embodiment of the present application is as follows: Figure 8 As shown, Figure 4 By comparison, it can be seen that the coherence of the edges is significantly improved, and is distributed between 0.5 and 1.

[0146] The fourth step of the PS-InSAR technical process is to estimate the linear deformation rate and residual elevation.

[0147] The linear deformation rate and elevation correction of PS points are estimated by least squares adjustment through the network model. Then, the least squares method is used to iteratively integrate these differential values ​​to the entire network edge to calculate the linear deformation rate and residual elevation value of each PS point.

[0148] The fifth step of the PS-InSAR technical process is to unwrap the residual phase.

[0149] Subtract the linear deformation rate and residual elevation phase components from the original phase to obtain a relatively clean residual phase. The residual phase of each interferogram is unwrapped in two dimensions using the minimum cost flow method to obtain a phase time series for each acquisition date of a set of SAR images. Let m be the number of SAR images processed, set the phase of the first scene image to zero, and obtain m-1 image phases.

[0150] Next, the atmospheric phase is estimated and removed. The unwrapped residual phase contains two main components, nonlinear deformation and atmospheric phase. By using the characteristics of high spatial correlation and low temporal correlation of atmospheric components for filtering, the atmospheric phase (APS) of the m-1 scene images is obtained. The average value of the atmospheric phase of the m-1 scene images is used to estimate the atmospheric phase of the first image. The obtained atmospheric phase is removed from the residual phase to obtain the nonlinear deformation phase.

[0151] Finally, the deformation rate is calculated. After removing the atmospheric phase and residual elevation from the original phase, the step iteration operation is repeated to obtain the linear deformation rate and the nonlinear deformation variable by unwrapping. The nonlinear deformation variable is converted into the nonlinear deformation rate and added to the linear deformation rate to obtain the overall deformation rate.

[0152] To summarize, the scheme of the embodiment of the present application, after the initial triangulated network is obtained by triangulating the network, the initial triangulated network is initially screened to achieve preliminary optimization of the initial triangulated network, and then the degree of network imbalance in the region is reduced through adaptive expansion and encryption processing. Finally, through coherence screening processing and connectivity screening processing, the PS points and edges with low coherence and connectivity in the encrypted network can be removed. It is suitable for mountainous areas where the deformation rate cannot be solved due to the undulating terrain, and overcomes the problem of losing measurement points (i.e., PS points) during the transfer unwrapping process due to the influence of phase incoherence caused by the undulating terrain, which helps to improve the effect of subsequent PS-InSAR.

[0153] Fig. 9 A schematic diagram of the structure of the adaptive networking device provided in the embodiment of the present application is shown in FIG. Fig. 9 As shown, the adaptive networking device includes:

[0154] A meshing module 91 is used to perform triangulation on a plurality of PS points on the interference pattern corresponding to the SAR image to obtain a corresponding initial triangulation network, wherein the initial triangulation network includes the plurality of PS points and a plurality of edges, and two vertices of any edge are the PS points;

[0155] A processing module 92 is used to perform initial screening processing on the initial triangulated network to obtain a seed network, wherein the seed network is composed of seed PS points and seed edges, wherein the seed PS points are a subset of the plurality of PS points, and the seed edges are a subset of the plurality of edges;

[0156] An encryption module 93, configured to perform adaptive extended encryption processing on the seed network according to the seed PS point and the seed edge to obtain an encrypted network;

[0157] The screening module 94 is used to perform coherence screening and connectivity screening on the encrypted network to obtain a target network.

[0158] In a possible implementation manner, the networking module 91 is specifically used to:

[0159] determining positions of the plurality of PS points on the interference pattern;

[0160] According to the positions of the multiple PS points on the interference graph, Delaunay triangulation is performed on the multiple PS points to obtain the initial triangulation network;

[0161] Among them, the initial triangulated network includes multiple triangles, and any triangle is composed of three PS points and edges connecting the three PS points in pairs; the circumscribed circle of any triangle does not contain any PS points other than the three PS points that constitute the triangle; after the diagonals of the convex quadrilateral formed by any two adjacent triangles are exchanged with each other, the minimum angle of the six internal angles of the convex quadrilateral no longer increases.

[0162] In a possible implementation manner, the processing module 92 is specifically configured to:

[0163] Executing a first operation, the first operation comprising: updating the PS points in the triangulated network updated for the i-1th time to obtain the triangulated network updated for the ith time;

[0164] Performing a second operation, the second operation comprising: determining the number of PS points in the triangulated network after the i-th update;

[0165] When the number of PS points in the triangulated network after the i-th update is less than the number of PS points in the triangulated network after the i-1-th update, update i to i+1, and repeat the first operation and the second operation; wherein, i is initially 1, i is a positive integer greater than or equal to 1, and the triangulated network after the 0th update is the initial triangulated network;

[0166] When the number of PS points in the triangulated mesh after the i-th update is equal to the number of PS points in the triangulated mesh after the (i-1)-th update, the triangulated mesh after the i-th update is determined as the seed mesh.

[0167] In a possible implementation manner, the processing module 92 is specifically configured to:

[0168] Obtaining the time coherence of each edge in the triangulated network after the (i-1)th update;

[0169] According to the temporal coherence and the first coherence threshold, removing low-quality edges in the triangulated network after the (i-1)th update, the low-quality edges being edges whose corresponding temporal coherence is less than or equal to the first coherence threshold;

[0170] According to the low-quality edge, isolated PS points in the triangulated network after the (i-1)th update are removed to obtain non-isolated PS points in the triangulated network after the (i-1)th update, wherein the isolated PS points are PS points that are not connected to any edge other than the low-quality edge;

[0171] Delaunay triangulation is performed according to the non-isolated PS points to obtain the triangulated network after the i-th update.

[0172] In a possible implementation manner, the encryption module 93 is specifically used for:

[0173] Performing statistical processing on the seed edges associated with each of the seed PS points to obtain the number of seed edges associated with each of the seed PS points in different quadrants;

[0174] Determine the quadrant to be encrypted of each seed PS point according to the number of seed edges associated with each seed PS point in different quadrants;

[0175] The seed network is adaptively extended and encrypted according to the seed PS points in the to-be-encrypted quadrant of each seed PS point to obtain the encrypted network.

[0176] In a possible implementation manner, the encryption module 93 is specifically used for:

[0177] For any seed PS point x, determine the PS points to be connected to the seed PS point x according to the distances between the seed PS point x and other PS points in the quadrant to be encrypted of the seed PS point x;

[0178] Determine the edge to be connected according to the seed PS point x and the PS point to be connected;

[0179] The seed network is adaptively extended and encrypted according to the to-be-connected edges corresponding to each of the seed PS points to obtain the encrypted network.

[0180] In a possible implementation, the screening module 94 is specifically used for:

[0181] Determining a first type of PS point and a second type of PS point in the encrypted network according to a second coherence threshold and a first connectivity threshold;

[0182] Acquire a skeleton network corresponding to the encrypted network according to the first type of PS points;

[0183] According to a third coherence threshold and a second connectivity threshold, the second type of PS points are connected to the skeleton network to obtain the target network, wherein the third coherence threshold is less than or equal to the second coherence threshold, and the second connectivity threshold is less than or equal to the first connectivity threshold.

[0184] The adaptive networking device provided in the embodiment of the present application is used to execute the above method embodiment, and its implementation principle and technical effect are similar, which will not be repeated in this embodiment.

[0185] Fig.10 An example of a physical structure diagram of an electronic device is shown in FIG. Fig.10 As shown, the electronic device may include: a processor 1010, a communication interface 1020, a memory 1030 and a communication bus 1040, wherein the processor 1010, the communication interface 1020 and the memory 1030 communicate with each other through the communication bus 1040. The processor 1010 may call the logic instructions in the memory 1030 to execute the adaptive network construction method, which includes: performing triangulation network construction on multiple PS points on the interference graph corresponding to the SAR image to obtain a corresponding initial triangulation network, wherein the initial triangulation network includes the multiple PS points and multiple edges, and the two vertices of any edge are the PS points; performing initial screening processing on the initial triangulation network to obtain a seed network, wherein the seed network is composed of seed PS points and seed edges, wherein the seed PS points are a subset of the multiple PS points, and the seed edges are a subset of the multiple edges; performing adaptive expansion encryption processing on the seed network according to the seed PS points and the seed edges to obtain an encrypted network; performing coherence screening processing and connectivity screening processing on the encrypted network to obtain a target network.

[0186] In addition, the logic instructions in the above-mentioned memory 1030 can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk.

[0187] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the adaptive meshing method provided by the above-mentioned methods, which includes: performing triangulation on multiple PS points on the interference pattern corresponding to the SAR image to obtain a corresponding initial triangulation network, wherein the initial triangulation network includes the multiple PS points and multiple edges, and the two vertices of any edge are the PS points; performing initial screening processing on the initial triangulation network to obtain a seed network, wherein the seed network is composed of seed PS points and seed edges, wherein the seed PS points are a subset of the multiple PS points, and the seed edges are a subset of the multiple edges; performing adaptive extended encryption processing on the seed network according to the seed PS points and the seed edges to obtain an encrypted network; performing coherence screening processing and connectivity screening processing on the encrypted network to obtain a target network.

[0188] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented by a processor to execute the adaptive meshing method provided by the above-mentioned methods when the computer program is executed, the method comprising: performing triangulation on a plurality of PS points on an interference pattern corresponding to a SAR image to obtain a corresponding initial triangulation, wherein the initial triangulation includes the plurality of PS points and a plurality of edges, and the two vertices of any edge are the PS points; performing initial screening processing on the initial triangulation to obtain a seed mesh, wherein the seed mesh is composed of seed PS points and seed edges, wherein the seed PS points are a subset of the plurality of PS points, and the seed edges are a subset of the plurality of edges; performing adaptive extended encryption processing on the seed mesh according to the seed PS points and the seed edges to obtain an encrypted mesh; performing coherence screening processing and connectivity screening processing on the encrypted mesh to obtain a target mesh.

[0189] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0190] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0191] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An adaptive networking method, characterized in that: include: Performing triangulation on a plurality of permanent scatterer PS points on an interference pattern corresponding to a synthetic aperture radar SAR image to obtain a corresponding initial triangulation network, wherein the initial triangulation network includes the plurality of PS points and a plurality of edges, and two vertices of any edge are both the PS points; Performing initial screening processing on the initial triangulated network to obtain a seed network, wherein the seed network is composed of seed PS points and seed edges, wherein the seed PS points are a subset of the multiple PS points, and the seed edges are a subset of the multiple edges; Performing adaptive extended encryption processing on the seed network according to the seed PS point and the seed edge to obtain an encrypted network; Performing coherence screening and connectivity screening on the encrypted network to obtain a target network; The step of performing adaptive extended encryption processing on the seed network according to the seed PS point and the seed edge to obtain an encrypted network includes: Performing statistical processing on the seed edges associated with each of the seed PS points to obtain the number of seed edges associated with each of the seed PS points in different quadrants; Determine the quadrant to be encrypted of each seed PS point according to the number of seed edges associated with each seed PS point in different quadrants; Performing adaptive extended encryption processing on the seed network according to the seed PS points in the to-be-encrypted quadrant of each seed PS point to obtain the encrypted network; The step of performing adaptive extended encryption processing on the seed network according to the seed PS points in the to-be-encrypted quadrant of each seed PS point to obtain the encrypted network comprises: For any seed PS point x, determine the PS points to be connected to the seed PS point x according to the distances between the seed PS point x and other PS points in the quadrant to be encrypted of the seed PS point x; Determine the edge to be connected according to the seed PS point x and the PS point to be connected; According to the to-be-connected edges corresponding to the seed PS points, the seed network is adaptively extended and encrypted to obtain the encrypted network; The step of performing coherence screening and connectivity screening on the encrypted network to obtain a target network includes: Determining a first type of PS point and a second type of PS point in the encrypted network according to a second coherence threshold and a first connectivity threshold; Acquire a skeleton network corresponding to the encrypted network according to the first type of PS points; According to a third coherence threshold and a second connectivity threshold, the second type of PS points are connected to the skeleton network to obtain the target network, wherein the third coherence threshold is less than or equal to the second coherence threshold, and the second connectivity threshold is less than or equal to the first connectivity threshold.

2. The adaptive networking method according to claim 1, characterized in that: The triangulating a plurality of permanent scatterer PS points on the interference pattern corresponding to the synthetic aperture radar SAR image to obtain a corresponding initial triangulated network includes: determining positions of the plurality of PS points on the interference pattern; According to the positions of the multiple PS points on the interference graph, Delaunay triangulation is performed on the multiple PS points to obtain the initial triangulation network; Among them, the initial triangulated network includes multiple triangles, and any triangle is composed of three PS points and edges connecting the three PS points in pairs; the circumscribed circle of any triangle does not contain any PS points other than the three PS points that constitute the triangle; after the diagonals of the convex quadrilateral formed by any two adjacent triangles are exchanged with each other, the minimum angle of the six internal angles of the convex quadrilateral no longer increases.

3. The adaptive networking method according to claim 2, characterized in that: The initial screening process is performed on the initial triangulated network to obtain a seed network, including: Executing a first operation, the first operation comprising: updating the PS points in the triangulated network updated for the i-1th time to obtain the triangulated network updated for the ith time; Performing a second operation, the second operation comprising: determining the number of PS points in the triangulated network after the i-th update; When the number of PS points in the triangulated network after the i-th update is less than the number of PS points in the triangulated network after the i-1-th update, update i to i+1, and repeat the first operation and the second operation; wherein, i is initially 1, i is a positive integer greater than or equal to 1, and the triangulated network after the 0th update is the initial triangulated network; When the number of PS points in the triangulated mesh after the i-th update is equal to the number of PS points in the triangulated mesh after the (i-1)-th update, the triangulated mesh after the i-th update is determined as the seed mesh.

4. The adaptive networking method according to claim 3, characterized in that: The updating process of the PS points in the triangulated network updated for the i-1th time to obtain the triangulated network updated for the ith time includes: Obtaining the time coherence of each edge in the triangulated network after the (i-1)th update; According to the temporal coherence and the first coherence threshold, removing low-quality edges in the triangulated network after the (i-1)th update, the low-quality edges being edges whose corresponding temporal coherence is less than or equal to the first coherence threshold; According to the low-quality edge, isolated PS points in the triangulated network after the (i-1)th update are removed to obtain non-isolated PS points in the triangulated network after the (i-1)th update, wherein the isolated PS points are PS points that are not connected to any edge other than the low-quality edge; Delaunay triangulation is performed according to the non-isolated PS points to obtain the triangulated network after the i-th update.

5. An adaptive networking device, characterized in that: include: A meshing module is used to triangulate the multiple PS points on the interference pattern corresponding to the SAR image to obtain a corresponding initial triangulated network, wherein the initial triangulated network includes the multiple PS points and multiple edges, and the two vertices of any edge are the PS points; A processing module, configured to perform initial screening processing on the initial triangulated network to obtain a seed network, wherein the seed network is composed of seed PS points and seed edges, wherein the seed PS points are a subset of the plurality of PS points, and the seed edges are a subset of the plurality of edges; An encryption module, used for performing adaptive extended encryption processing on the seed network according to the seed PS point and the seed edge to obtain an encrypted network; A screening module, used for performing coherence screening and connectivity screening on the encrypted network to obtain a target network; The step of performing adaptive extended encryption processing on the seed network according to the seed PS point and the seed edge to obtain an encrypted network includes: Performing statistical processing on the seed edges associated with each of the seed PS points to obtain the number of seed edges associated with each of the seed PS points in different quadrants; Determine the quadrant to be encrypted of each seed PS point according to the number of seed edges associated with each seed PS point in different quadrants; Performing adaptive extended encryption processing on the seed network according to the seed PS points in the to-be-encrypted quadrant of each seed PS point to obtain the encrypted network; The step of performing adaptive extended encryption processing on the seed network according to the seed PS points in the to-be-encrypted quadrant of each seed PS point to obtain the encrypted network comprises: For any seed PS point x, determine the PS points to be connected to the seed PS point x according to the distances between the seed PS point x and other PS points in the quadrant to be encrypted of the seed PS point x; Determine the edge to be connected according to the seed PS point x and the PS point to be connected; According to the to-be-connected edges corresponding to the seed PS points, the seed network is adaptively extended and encrypted to obtain the encrypted network; The step of performing coherence screening and connectivity screening on the encrypted network to obtain a target network includes: Determining a first type of PS point and a second type of PS point in the encrypted network according to a second coherence threshold and a first connectivity threshold; Acquire a skeleton network corresponding to the encrypted network according to the first type of PS points; According to a third coherence threshold and a second connectivity threshold, the second type of PS points are connected to the skeleton network to obtain the target network, wherein the third coherence threshold is less than or equal to the second coherence threshold, and the second connectivity threshold is less than or equal to the first connectivity threshold.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the adaptive networking method according to any one of claims 1 to 4 is implemented.

7. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the adaptive networking method according to any one of claims 1 to 4 is implemented.

8. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the adaptive networking method according to any one of claims 1 to 4 is implemented.

Images