Coherence proxy-based big data InSAR interference pair optimization method, medium and equipment

Abstract

The invention discloses a big data InSAR (Interferometric Synthetic Aperture Radar) interference pair optimization method based on coherence agents, which comprises the following steps of: calculating a coherence graph of a sampling interference pair, and calculating a time baseline coherence agent, a space baseline coherence agent, a seasonal coherence agent and a comprehensive coherence agent; selecting a plurality of proportional interference pairs with the highest coherence agents as additional redundant interference pairs, and taking a union set of the redundant interference pairs and the interference pair network of the minimum spanning tree to obtain a high-coherence agent interference pair network; and detecting whether all interference pairs in the high-coherence proxy interference pair network can form phase triangular closure or not, and performing phase triangular closure complementation on the interference pairs which do not form the phase triangular closure to obtain a final InSAR preferred interference pair network. According to the optimal selection method, the phase unwrapping and deformation estimation precision is improved, the number of interference pairs needing to be calculated under the big data InSAR condition is greatly reduced, the calculation burden is relieved, and support is provided for implementation of InSAR near-real-time processing.

Description

Technical Field

[0001] The present invention relates to the field of synthetic aperture radar interferometry technology in the technical field of time series data processing, and in particular to a method, medium and device for optimizing big data InSAR interferometry based on coherence agent. Background Art

[0002] Interferometric pair optimization is an important step in the interferometric synthetic apertureradar (InSAR) time series processing flow, and plays a key role in improving the InSAR deformation monitoring accuracy and spatial coverage density. Conventional interferometric pair optimization methods empirically set the temporal baseline threshold and spatial baseline threshold to eliminate pairs with significant temporal and spatial incoherence, which not only reduces the adverse effects of low-coherence interferometric pairs on phase unwrapping results and deformation estimation accuracy, but also greatly reduces the computational cost.

[0003] However, Ansari et al. found that using only short-baseline interferometric pairs would introduce fading signals, and this strategy did not consider the impact of seasonal incoherence. Therefore, some scholars proposed to directly calculate the average coherence or effective coherence ratio of all available interferometric pairs for interferometric pair optimization, which proved that deformation accuracy and spatial coverage density could be improved. Some scholars also combined spectral clustering, all-pair shortest path algorithm, minimum spanning tree and quadtree algorithm to optimize SAR interferometric pairs. However, the number of available interferometric pairs and the number of SAR data show a power growth relationship, that is, N SAR images covering the same area can generate N(N□1) / 2 interferometric images. In the era of InSAR big data, SAR data is growing rapidly, and the number of interferometric pairs is also exponentially growing. It is difficult to calculate the coherence indicators of all available interferometric pairs, and the high computational burden will hinder the realization of InSAR near-real-time processing. Summary of the invention

[0004] The purpose of the present invention is to provide a method, medium and device for optimizing InSAR interference based on a big data based on a coherence agent in view of the problems existing in the prior art.

[0005] To achieve the above object, the technical solution adopted by the present invention is: A method for optimizing InSAR interferometry pairs based on big data based on coherence proxy includes the following steps: According to the SAR images covering the same area, randomly sampling from all the interferometric pairs that can be generated, obtaining a number of sampled interferometric pairs, and calculating the coherence graph of the sampled interferometric pairs; According to the coherence map, a temporal baseline coherence proxy, a spatial baseline coherence proxy, a seasonal coherence proxy and a comprehensive coherence proxy are calculated; According to the temporal baseline coherence proxy, the spatial baseline coherence proxy, the seasonal coherence proxy and the comprehensive coherence proxy, selecting a number of proportional interference pairs with the highest coherence proxy as additional redundant interference pairs; Calculating the interference pair network of the minimum spanning tree, taking the union of the redundant interference pairs and the interference pair network of the minimum spanning tree, and obtaining a high-coherence proxy interference pair network; Whether all the interference pairs in the high coherence proxy interference pair network can form phase triangle closure is detected, and the phase triangle closure is supplemented for the interference pairs that have not formed phase triangle closure, so as to obtain the final InSAR preferred interference pair network.

[0006] This big data InSAR interferometer pair optimization method based on coherence agent not only improves the phase unwrapping and deformation estimation accuracy, but also greatly reduces the number of interferometer pairs that need to be calculated in the case of big data InSAR, reduces the computational burden, and provides support for the realization of InSAR near-real-time processing.

[0007] During the data processing process, this preferred method fully considers the time baseline coherence proxy, the space baseline coherence proxy, the seasonal coherence proxy and the comprehensive coherence proxy, and can select a certain proportion of interference pairs with the highest coherence proxy. Combined with the minimum spanning tree algorithm, it avoids the loss of accuracy caused by too many independent subsets.

[0008] This optimization method can further detect whether the obtained high-coherence proxy interference pair network can form a phase triangle closure, and can also complement the interference pairs that have not formed a phase triangle closure, providing support for phase unwrapping error correction. The obtained interference pair network is the most preferred in the case of big data, which greatly improves the InSAR deformation monitoring accuracy and spatial coverage density.

[0009] Furthermore, the calculation method of the coherence graph is as follows: , In the formula, i and j represents the pixel position in the coherence map, γ(i, j) Indicates that in the pixel (i, j) The coherence coefficient at S 1 (i, j) and S 2 (i, j) The reference image and the secondary image are respectively in the pixel ( i , j ), S* represents a conjugate complex number, such as S 2 * (i, j) for S 2 (i, j) The complex conjugate of m and n is the size of the coherence window. m and n can be positive integers. |·| represents the absolute value of a complex number, for example | S 1 (i, j) |For S 1 (i, j) The absolute value of .

[0010] Furthermore, the calculation method of the time baseline coherence proxy is as follows: The average coherence of the interference pairs with the same time baseline is calculated, and the average coherence is used as the dependent variable, and the time baseline is used as the independent variable. The functional relationship between the average coherence and the time baseline is fitted using the following formula as a proxy for the time baseline coherence: , In the formula, c time is the average coherence of the interferometric pairs with equal time baselines, B T is the time baseline, c 0 t and c ∞ t denote the short-term coherence and long-term coherence related to the time baseline, respectively. t t represents the temporal decorrelation rate, A t 、oh t ,f t They represent the amplitude, angular frequency and phase of the periodic variation of the baseline over time, respectively. |·| represents the absolute value, for example | B T |For B T The absolute value of .

[0011] Furthermore, the calculation method of the spatial baseline coherence proxy is as follows: The spatial baseline interval is divided into several meters, and the average coherence of the interference pairs in the same spatial baseline interval is calculated and used as the dependent variable. The spatial baseline is used as the independent variable, and the functional relationship between the two is fitted using the following formula as a proxy for the spatial baseline coherence: , In the formula, c spatial is the average coherence in the same spatial baseline interval, B P is the spatial baseline, c 0 P and c ∞ P denote the short-term coherence and long-term coherence related to the spatial baseline, respectively. t P represents the spatial decorrelation rate.

[0012] Furthermore, the seasonal coherence proxy is calculated as follows: For a certain date, the average coherence of all interference pairs whose main or auxiliary image dates are the same as the certain date is calculated and used as the dependent variable. The date is used as the independent variable, and the functional relationship between the two is fitted using the following formula: , In the formula, c P&S is the average coherence of images with the same image date, day P&S is the date difference between the main image or auxiliary image constituting the interferometric pair and the first SAR image. A S 、the S 、e S Respectively represent the amplitude, horizontal offset and vertical offset of the cosine function; Then use the following formula as a proxy for seasonal coherence: , In the formula, c seasonal is a proxy for seasonal coherence, day P and day S It is the date of the primary and secondary images that make up the interferometric pair relative to the first SAR image.

[0013] Furthermore, the calculation method of the comprehensive coherence proxy is as follows: the product of the time baseline coherence proxy, the space baseline coherence proxy and the seasonal coherence proxy is taken as the comprehensive coherence proxy.

[0014] Furthermore, the interference pair network of the minimum spanning tree is calculated by taking the comprehensive coherence agent as the weight of the edge, taking the SAR image and the interference pair as the points and edges in the graph theory, and using the minimum spanning tree algorithm to obtain the interference pair network of the minimum spanning tree.

[0015] Further, the method of detecting and completing the phase triangle closure is as follows: For any interference pair to be detected, all dates except the two dates of the main image and auxiliary image of the interference pair are traversed. If there is any date that forms an interference pair with the two dates of the main image and auxiliary image of the interference pair in the current interference pair network, then the interference pair is considered to have phase triangle closure, otherwise it does not have phase triangle closure; for interference pairs without phase triangle closure, all dates except the two dates of the main image and auxiliary image of the interference pair are traversed, and the average coherence proxy of the interference pair formed with the two dates of the main image and auxiliary image of the interference pair is calculated, the highest value is taken and the corresponding interference pair is added to the interference pair network.

[0016] The present invention also provides a medium, which is a computer-readable storage medium, wherein the computer-readable storage medium includes a stored program, and when the program is executed by a processor, the above-mentioned big data InSAR interference optimization method based on coherence agent is implemented.

[0017] The present invention also provides a device, which is an electronic device, comprising at least one processor and at least one memory connected to the processor; wherein the processor is used to call program instructions in the memory to execute the coherence agent-based big data InSAR interference pair optimization method as described above.

[0018] Compared with the prior art, the beneficial effects of the present invention are: 1. This big data InSAR interference pair optimization method based on coherence proxy not only improves the phase unwrapping and deformation estimation accuracy, but also greatly reduces the number of interference pairs that need to be calculated in the case of big data InSAR, reduces the calculation burden, and provides support for the realization of InSAR near-real-time processing; 2. In the data processing process, this optimization method fully considers the time baseline coherence proxy, the space baseline coherence proxy, the seasonal coherence proxy and the comprehensive coherence proxy, and can select a certain proportion of interference pairs with the highest coherence proxy. Combined with the minimum spanning tree algorithm, it avoids the accuracy loss caused by too many independent subsets; 3. This optimization method can further detect whether the obtained high coherence proxy interference pair network can form a phase triangle closure, and can also complement the interference pairs that have not formed a phase triangle closure, providing support for phase unwrapping error correction. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1It is a flow chart of a method for optimizing the InSAR interferometry based on a big data based on a coherence agent of the present invention; Figure 2 The calculation results of the coherence coefficient of the present invention are as follows: (a) is an example diagram of the time baseline coherence coefficient, (b) is an example diagram of the space baseline coherence coefficient, (c) is an example diagram of the seasonal coherence coefficient, and (d) is an example diagram of the comprehensive coherence coefficient; Figure 3 Schematic diagram of the results of the intermediate process of interference graph optimization: (a) is the minimum spanning tree interference pair network, (b) is the high coherence proxy interference pair network; Figure 4 Schematic diagram of interference pattern optimization results: (a) is the high coherence agent combined triangle phase closure interference network of the present invention, and (b) is the space-time baseline threshold method interference network. DETAILED DESCRIPTION

[0020] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings of the present invention. Obviously, the described embodiments are only 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.

[0021] like Figure 1 As shown, a method for optimizing InSAR interferometry pairs based on a big data coherence agent includes the following steps: Step 1: Based on N SAR images covering the same area, a certain proportion of interference pairs are randomly sampled from all the interference pairs that can be generated to obtain M interference pairs, and the coherence graphs of these M interference pairs are calculated. The calculation method of the coherence graph is as follows: , In the formula, i and j represents the pixel position in the coherence map, S 1 (i, j) and S 2 (i, j) The reference image and the secondary image are in the pixel ( i , j ), S * represents the conjugate complex number, m and n is the size of the coherence window for calculation, and |·| represents the absolute value of the complex number.

[0022] Step 2: Based on the coherence graph, calculate the temporal baseline coherence proxy, the spatial baseline coherence proxy, the seasonal coherence proxy and the comprehensive coherence proxy.

[0023] The calculation method of the time baseline coherence proxy is as follows: The average coherence of the interference pairs with the same time baseline is calculated, and the average coherence is used as the dependent variable, and the time baseline is used as the independent variable. The functional relationship between the average coherence and the time baseline is fitted using the following formula as a proxy for the time baseline coherence: , In the formula, c time is the average coherence of interferometric pairs with equal time baselines, i.e., a proxy for time baseline coherence, B T is the time baseline, c 0 t and c ∞ t denote the short-term coherence and long-term coherence related to the time baseline, respectively. t t represents the temporal decorrelation rate, A t 、oh t ,f t They represent the amplitude, angular frequency, and phase of the periodic variation of the baseline over time, respectively, and |·| represents the absolute value.

[0024] The calculation method of the spatial baseline coherence proxy is as follows: The spatial baseline interval is divided into 5-meter intervals, and the average coherence of the interference pairs in the same spatial baseline interval is calculated and used as the dependent variable. The spatial baseline is used as the independent variable, and the functional relationship between the two is fitted using the following formula as a proxy for spatial baseline coherence: , In the formula, c spatial is the average coherence in the same spatial baseline interval, i.e., the spatial baseline coherence proxy, B P is the spatial baseline, c 0 P and c ∞ P denote the short-term coherence and long-term coherence related to the spatial baseline, respectively. t P represents the spatial decorrelation rate.

[0025] The seasonal coherence proxy is calculated as follows: For a certain date, the average coherence of all interference pairs whose main or auxiliary image dates are the same as the certain date is calculated and used as the dependent variable. The date is used as the independent variable, and the functional relationship between the two is fitted using the following formula: , In the formula, c P&S is the average coherence of images with the same image date, day P&S is the date difference between the main image or auxiliary image constituting the interferometric pair and the first SAR image. A S 、the S 、e S Respectively represent the amplitude, horizontal offset and vertical offset of the cosine function; Then use the following formula as a proxy for seasonal coherence: , In the formula, c seasonal is a proxy for seasonal coherence, day P and day S is the date of the primary and secondary images constituting the interferometric pair relative to the first SAR image, A S 、the S 、e S All are utilization Parameters to calculate.

[0026] The calculation method of the comprehensive coherence proxy is as follows: the product of the temporal baseline coherence proxy, the spatial baseline coherence proxy and the seasonal coherence proxy is used as a comprehensive coherence proxy as a substitute indicator of coherence, as shown in the following formula: .

[0027] In this embodiment, the calculation results of the time baseline coherence proxy, the space baseline coherence proxy, the seasonal coherence proxy and the comprehensive coherence proxy are as follows: Figure 2 shown.

[0028] Step 3: Using the comprehensive coherence agent as the edge weight, and the SAR image and interference pair as the points and edges in the graph theory, the minimum spanning tree algorithm is used to obtain the interference pair network of the minimum spanning tree, such as Figure 3 (a) as shown; According to the temporal baseline coherence proxy, the spatial baseline coherence proxy, the seasonal coherence proxy and the comprehensive coherence proxy, a comprehensive coherence proxy threshold is obtained, that is, a number of interference pairs with the highest coherence proxy ratio are selected as additional redundant interference pairs; the redundant interference pairs are combined with the interference pair network of the minimum spanning tree to obtain a high coherence proxy interference pair network, such as Figure 3 (b) shown.

[0029] Step 4: Check whether all the interference pairs in the interference pair network obtained in step 3 can form phase triangle closure, and complete the phase triangle closure for the interference pairs that have not formed phase triangle closure to obtain the final InSAR optimal interference pair network, as shown in Figure 4 (a) is shown; the specific process of phase triangle closure detection and completion is as follows: for any interference pair to be detected, all dates except the two dates of the main and auxiliary images of the interference pair are traversed. If there is any date that forms an interference pair with the two dates of the main and auxiliary images of the interference pair in the current interference pair network, then the interference pair is considered to have phase triangle closure, otherwise it does not have phase triangle closure. For interference pairs that do not have phase triangle closure, all dates except the two dates of the main and auxiliary images of the interference pair are traversed, and the average coherence proxy of the interference pair formed with the two dates of the main and auxiliary images of the interference pair is calculated. The highest value is taken and the corresponding interference pair is added to the interference pair network.

[0030] As a comparative method of the present invention, the interference effect of the spatial-temporal baseline threshold method on the network is as follows: Figure 4 (b) shown.

[0031] This big data InSAR interferometer pair optimization method based on coherence agent not only improves the phase unwrapping and deformation estimation accuracy, but also greatly reduces the number of interferometer pairs that need to be calculated in the case of big data InSAR, reduces the computational burden, and provides support for the realization of InSAR near-real-time processing.

[0032] In the data processing process, this preferred method fully considers the time baseline coherence proxy, the space baseline coherence proxy, the seasonal coherence proxy and the comprehensive coherence proxy, and can select a certain proportion of interference pairs with the highest coherence proxy, and combines the minimum spanning tree algorithm to avoid the decrease in accuracy caused by too many independent subsets. Combined with the detection and completion of phase delta closure, it provides support for phase unwrapping error correction.

[0033] In specific implementation, the above process can be realized by automatic operation of the process using computer software (program) technology.

[0034] Specifically, the present invention also provides a medium, which is a computer-readable storage medium, and the computer-readable storage medium includes a stored program, and when the program is executed by a processor, the big data InSAR interference pair optimization method based on coherence agent as described above is implemented, and steps 1 to 4 are executed.

[0035] The present invention also provides a device, which is an electronic device, comprising at least one processor and at least one memory connected to the processor; wherein the processor is used to call program instructions in the memory, execute the big data InSAR interference pair optimization method based on coherence agent as described above, and execute steps 1 to 4.

[0036] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on the computer, the process or function according to the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website site, a computer, a server or a data center to another website site, a computer, a server or a data center by wired or wireless means. The computer-readable storage medium can be any available medium that a computer can store or a data storage device such as a server or a data center that includes one or more available media integration. The available medium can be a magnetic medium, an optical medium, or a semiconductor medium, etc.

[0037] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A big data InSAR interferometric pair optimization method based on coherence proxy, characterized in that: The steps include: According to the SAR images covering the same area, randomly sampling from all the interferometric pairs that can be generated, obtaining a number of sampled interferometric pairs, and calculating the coherence graph of the sampled interferometric pairs; According to the coherence map, a temporal baseline coherence proxy, a spatial baseline coherence proxy, a seasonal coherence proxy and a comprehensive coherence proxy are calculated; According to the temporal baseline coherence proxy, the spatial baseline coherence proxy, the seasonal coherence proxy and the comprehensive coherence proxy, selecting a number of proportional interference pairs with the highest coherence proxy as additional redundant interference pairs; Calculating the interference pair network of the minimum spanning tree, taking the union of the redundant interference pairs and the interference pair network of the minimum spanning tree, and obtaining a high-coherence proxy interference pair network; Whether all the interference pairs in the high coherence proxy interference pair network can form phase triangle closure is detected, and the phase triangle closure is supplemented for the interference pairs that have not formed phase triangle closure, so as to obtain the final InSAR preferred interference pair network.

2. The method for optimizing InSAR interferometry based on coherence agent based on big data according to claim 1, characterized in that: The coherence graph is calculated as follows: ， In the formula, i and j represents the pixel position in the coherence map, γ(i, j) Indicates that in the pixel (i, j) The coherence coefficient at S 1 (i, j) and S 2 (i, j) The reference image and the secondary image are respectively in the pixel ( i , j ), S * represents the conjugate complex number, m and n is the size of the coherence window for computing coherence.

3. The method for optimizing InSAR interferometry based on coherence proxy based on big data according to claim 1, characterized in that: The temporal baseline coherence proxy is calculated as follows: The average coherence of the interference pairs with the same time baseline is calculated, and the average coherence is used as the dependent variable, and the time baseline is used as the independent variable. The functional relationship between the average coherence and the time baseline is fitted using the following formula as a proxy for the time baseline coherence: ， In the formula, γ time is the average coherence of the interferometric pairs with equal time baselines, B T is the time baseline, γ 0 t and γ ∞ t denote the short-term coherence and long-term coherence related to the time baseline, respectively. τ t represents the temporal decorrelation rate, A t ,ω t ,φ t They represent the amplitude, angular frequency and phase of the periodic variation of the time baseline respectively.

4. The method for optimizing InSAR interferometry based on coherence agent based on big data according to claim 1, characterized in that: The calculation method of the spatial baseline coherence proxy is as follows: The spatial baseline interval is divided into several meters, and the average coherence of the interference pairs in the same spatial baseline interval is calculated and used as the dependent variable. The spatial baseline is used as the independent variable, and the functional relationship between the two is fitted using the following formula as a proxy for the spatial baseline coherence: ， In the formula, γ spatial is the average coherence in the same spatial baseline interval, B P is the spatial baseline, γ 0 P and γ ∞ P denote the short-term coherence and long-term coherence related to the spatial baseline, respectively. τ P represents the spatial decorrelation rate.

5. The method for optimizing InSAR interferometry based on coherence agent based on big data according to claim 1, characterized in that: The seasonal coherence proxy is calculated as follows: For a certain date, the average coherence of all interference pairs whose main or auxiliary image dates are the same as the certain date is calculated and used as the dependent variable. The date is used as the independent variable, and the functional relationship between the two is fitted using the following formula: ， In the formula, γ P&S is the average coherence of images with the same image date, day P&S is the date difference between the main image or auxiliary image constituting the interferometric pair and the first SAR image. A S ,θ S , ε S Respectively represent the amplitude, horizontal offset and vertical offset of the cosine function; Then use the following formula as a proxy for seasonal coherence: ， In the formula, γ seasonal is a proxy for seasonal coherence, day P and day S It is the date of the primary and secondary images that make up the interferometric pair relative to the first SAR image.

6. The method for optimizing InSAR interferometry based on coherence agent based on big data according to claim 1, characterized in that: The calculation method of the comprehensive coherence proxy is as follows: the product of the time baseline coherence proxy, the space baseline coherence proxy and the seasonal coherence proxy is taken as the comprehensive coherence proxy.

7. The method for optimizing InSAR interferometry based on big data based on coherence proxy according to claim 1, characterized in that: The calculation method of the interference pair network of the minimum spanning tree is: taking the comprehensive coherence agent as the weight of the edge, taking the SAR image and the interference pair as the points and edges in the graph theory, and using the minimum spanning tree algorithm to obtain the interference pair network of the minimum spanning tree.

8. The method for optimizing InSAR interferometry based on big data based on coherence proxy according to claim 1, characterized in that: The method for detecting and completing the phase triangle closure is as follows: For any interference pair to be detected, all dates except the two dates of the main image and auxiliary image of the interference pair are traversed. If there is any date that forms an interference pair with the two dates of the main image and auxiliary image of the interference pair in the current interference pair network, then the interference pair is considered to have phase triangle closure, otherwise it does not have phase triangle closure; for interference pairs without phase triangle closure, all dates except the two dates of the main image and auxiliary image of the interference pair are traversed, and the average coherence proxy of the interference pair formed with the two dates of the main image and auxiliary image of the interference pair is calculated, the highest value is taken and the corresponding interference pair is added to the interference pair network.

9. A medium, which is a computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, and when the program is executed by a processor, the big data InSAR interferometry optimization method based on coherence agent according to any one of claims 1 to 8 is implemented.

10. A device, which is an electronic device, characterized in that: The electronic device includes at least one processor and at least one memory connected to the processor; wherein the processor is used to call program instructions in the memory to execute the big data InSAR interference pair optimization method based on coherence agent as described in any one of claims 1-8.

Images