CN115825955A - A Polarimetric Time-Series InSAR Method Based on Coherent Matrix Adaptive Decomposition - Google Patents

Abstract

The invention discloses a polarization time sequence InSAR method based on coherent matrix self-adaptive decomposition, which comprises the steps of obtaining a research time sequence multi-polarization time sequence SAR image, preprocessing the SAR image and identifying homogeneous pixels; carrying out polarization optimization under different criteria according to PS and DS targets; calculating time sequence coherence by using the interference phase subjected to polarization optimization and decomposed by characteristic values, selecting high-quality pixels, performing polarization time sequence interference measurement by using a differential interference pattern subjected to polarization optimization, obtaining deformation monitoring points, and finishing deformation monitoring on a research area; according to the method, the PS target and the DS target are distinguished through homogeneous pixel identification, an optimal scattering mechanism is searched by utilizing eigenvalue decomposition and BEST algorithm, and different strategies are applied to the PS target and the DS target for optimization, so that the phase quality of an interference pattern is obviously improved, the phase noise of the interference pattern is reduced, and interference fringes are clearer.

Description

A Polarimetric Time-Series InSAR Method Based on Coherent Matrix Adaptive Decomposition

technical field

The invention relates to the technical field of surface deformation monitoring, in particular to a polarization time-series InSAR method based on coherent matrix adaptive decomposition.

Background technique

Surface deformation includes ground subsidence and surface subsidence. Traditional surface deformation monitoring technologies and means mainly include: leveling technology, GPS measurement technology, electronic rangefinder measurement, and underground bedrock mark and layered mark measurement technology. However, these "contact" monitoring techniques require a lot of manpower, material and financial resources, and usually require surveyors to enter the monitoring area, which increases the difficulty of monitoring work and even poses certain safety hazards. In addition, these traditional "point"-based monitoring technologies and means have disadvantages such as low spatial resolution of monitoring results, long monitoring cycle, small monitoring range and high monitoring costs.

The rapid development of Synthetic Aperture Radar (SAR) system has provided abundant data sources for earth observation. And it has been widely used in deformation monitoring. The traditional PS-InSAR technology can obtain better results in urban areas with artificial structures such as buildings, bridges, and dams by selecting permanent scattering targets (Persistent Scatterer, PS), while for vegetation, bare soil, deserts, and roads For these distributed scatter targets (Distributed Scatter, DS) targets with complex and changeable scattering characteristics, low echo signal strength, and weak backscattering ability, the effect obtained by PS-InSAR technology is often not ideal. As far as the actual situation is concerned, the surface features are complex and diverse, and PS targets and DS targets usually exist on the ground at the same time. In the prior art, Polarimetric Interferometry SAR (PolInSAR) was proposed. Compared with single-polarization data to monitor surface deformation, multi-polarization data can make full use of the different responses of different scatterers to polarization signals. , improve the interferometric coherence, optimize the quality of the interferometric phase, increase the density of monitoring points, and improve the quality of monitoring points.

Contents of the invention

Aiming at the above-mentioned technical deficiencies, the purpose of the present invention is to provide a polarization time-series InSAR method based on coherent matrix adaptive decomposition, which first distinguishes PS and DS targets through homogeneous pixel identification, and utilizes eigenvalue decomposition and BEST algorithm, Search for the optimal scattering mechanism and apply different strategies to PS and DS targets for optimization. The phase quality of the interferogram is significantly improved, the phase noise of the interferogram is reduced, and the interference fringes are clearer.

In order to solve the problems of the technologies described above, the present invention adopts the following technical solutions:

The present invention provides a polarization timing InSAR method based on coherent matrix adaptive decomposition, including the following methods:

S1. Obtain and study multi-polarization time-series SAR images in the monitoring area, preprocess the SAR images and identify homogeneous pixels, and obtain PS and DS targets;

S2. Carry out polarization optimization under different criteria according to the PS and DS objectives, including the following steps:

S21. Under the Pauli basis, construct a polarization scattering vector for the multi-polarization data;

S22. Constructing the PS target polarization coherence matrix T ^PS and the polarization vector interferometric phase of the PS target;

S23. Perform polarization-adaptive MMSE filtering on the polarization coherence matrix T ^DS constructed by the DS target;

S24, polarization coherence matrix eigenvalue decomposition;

S25. For the PS and DS targets, respectively adopt the D _A criterion and the γ criterion to perform polarization optimization;

S3: Use the polarization-optimized interferometric phase of eigenvalue decomposition to calculate the time-series coherence, select high-quality pixels, use the polarization-optimized differential interferogram to perform polarization time-series interferometry, and obtain deformation monitoring points, complete Deformation monitoring of the study area.

Preferably, the specific method of step S1 is as follows: after obtaining and studying the multi-polarization time-series SAR images in the monitoring area, the SAR images are pre-registered, cropped, and geocoded, and the intensity information of the multi-polarization time-series SAR images is used, based on The homogeneous pixel identification method of statistical test divides the multi-polarization SAR image into PS and DS targets under their respective polarizations, and then uses the amplitude deviation D _A criterion to transform the pixels with a D _A value less than 0.33 in the DS target For the PS target, since different ground objects have different intensity information in different polarizations, after identifying homogeneous pixels in multi-polarization SAR images, the results of homogeneous point identification in different polarizations are combined to obtain the final PS and DS target

Preferably, the specific method of step S21 is as follows:

The polarimetric radar transmits and receives electromagnetic pulse signals in two different ways, vertical and horizontal, and records images of different polarizations according to the way the antenna sends and receives signals during imaging. There are four polarization modes of VV, VH, HH, and HV. The data constructs the scattering matrix S:

where S represents the scattering matrix, S _hh represents the single-view complex SAR image under HH polarization, S _vv represents the single-view complex SAR image under VV polarization, S _vh represents the single-view complex SAR image under VH polarization, _Shv represents the single-view complex SAR image under HV polarization;

For fully polarized data, construct the scattering vector k _q under the Pauli basis as follows:

In the formula, q represents full polarization, according to the reciprocity theorem S _vh =S _hv , so only S _vh represents cross polarization, the same below, and T represents matrix transposition operation;

For dual-polarization data, there are several ways to combine different polarizations:

For the SAR image data combined by HH and VV, the scattering vector k _d under the Pauli basis is constructed as follows:

For dual-polarization SAR image data combined with a co-polarization (HH, VV) and a cross-polarization (VH, HV), the scattering vector k _d under the Pauli basis is constructed as follows:

k _d =[S _xx , 2S _vh ] ^T

In the formula, d represents dual polarization; S _xx represents the single-view complex SAR image under HH or VV polarization;

Polarization vector interferometry is different from single polarization scalar interferometry. Single polarization scalar interferometry is the conjugate multiplication between the first and second images of the registration, while polarization vector interferometry needs to construct the respective scattering of the two polarized images Vector, and introduce the scattering mechanism to do the outer product to obtain the interference phase, and the scattering mechanism is represented by the unit complex projection vector ω;

For fully polarized data, ω _q is given by:

where a, β and ψ denote the polarization parameters of the scattering mechanism vector, j and e denote the imaginary unit and natural constant, respectively;

For dual-polarization data, ω _d is given by:

In order to carry out polarization vector interference, the unit complex vector ω _i of the specific scattering mechanism of the two images is used to project the scattering vector _ki of the first image and the second image respectively by the following formula to obtain μ _i :

μ _i =ω _i ^H ·k _i , i=1, 2

In the formula, i represents the i-th image, H represents the conjugate transpose, and μ _i represent the components of their specific scattering mechanisms, similar to single-view complex images in single polarization.

Preferably, the specific method of step S22 is as follows:

The coherence matrix can effectively reduce the influence of speckle noise, while the spatial average coherence matrix loses the spatial resolution of the image. Therefore, in order to preserve the spatial resolution, this study adopts the temporal coherence matrix and directly constructs the polarization coherence matrix for the PS target:

where k _j represents the jth polarized scattering vector;

Generate the polarization vector interferophase based on two complex scalars:

I=μ ₁ ·μ ₂ ^*

where * is the complex conjugate operator.

Preferably, the specific method of step S23 is as follows:

DS targets are mostly homogeneous objects such as vegetation, bare soil, roads, etc. The echo signal intensity is low, the backscattering ability is weak, and the phase stability is poor. Therefore, homogeneity filtering is required for DS targets after identifying homogeneous pixels. , to improve the phase quality of DS targets; there are many homogeneous pixel recognition algorithms based on statistical tests: BWS hypothesis test, two-sample t hypothesis test, generalized likelihood ratio hypothesis test, FaSHPS test algorithm and HTCI algorithm, etc.;

Construct the polarization coherence matrix T ^DS for the DS target:

where k ₁ and k ₂ represent the scattering vectors under the Pauli basis, T ₁₁ and T ₂₂ represent the coherence matrix, Ω ₁₂ represents the polarization coherence matrix, in the MMSE polarization filtering of homogeneous pixels, it is first necessary to calculate the polarization coherence The span image of matrix T:

span=Trace(T ₁₁ )+Trace(T ₂₂ )

In the formula, Trace represents the trace of the matrix, and the filter weight coefficient b is calculated by the value of span as follows:

In the formula, SHPS represents a homogeneous pixel, and the coef value is 0.51. Finally, the MMSE filtering of the homogeneous pixel is completed by the following formula, and the MMSE filtering is completed for the polarization coherence matrix T ^DS of the DS target:

T ^DS ＝T _mean +b(T _ori -T _mean )

In the formula, T _mean represents the average coherence matrix of homogeneous pixels in the window, and T _ori represents the original coherence matrix.

Preferably, the specific method of step S24 is as follows:

In this step, the polarization coherence matrices of the ^TPS and ^TDS targets can be combined into one for eigenvalue decomposition, or can be separately eigenvalue decomposed. The eigenvalue decomposition of the coherence matrix T is performed; the time-series average polarization coherence matrix can be decomposed as:

表示平均时间极化相干矩阵，λ_i表示特征值，u_i表示特征向量，num表示极化通道个数；In the formula

Represents the average time polarization coherence matrix, λ _i represents the eigenvalue, u _i represents the eigenvector, and num represents the number of polarization channels;

When fully polarized data is used, num=3, three eigenvalues λ ₁ ≥ λ ₂ ≥ _{λ 3} ≥ 0 and corresponding eigenvectors u ₁ , u ₂ , u ₃ are obtained;

When dual-polarization data is used, num=2, two eigenvalues λ ₁ ≥λ ₂ ≥0 and corresponding eigenvectors u ₁ , u ₂ are obtained.

Preferably, the specific method of step S24 is as follows:

Step 1, _DA criterion optimization of PS target:

The phase quality evaluation index D _A is obtained by comparing the standard deviation of the time-series SAR image amplitude with the mean value, which is used to optimize the PS target. The smaller the D _A value, the better the phase quality, which is obtained by the following formula:

表示Pol极化下幅度的标准差，

表示Pol极化下幅度的均值，其中Pol是VV、HH或VH极化通道，对于极化时序影像，不同散射体机制ω下的D_A通过下式得到：In the formula

Indicates the standard deviation of the amplitude under the Pol polarization,

Indicates the mean value of the amplitude under Pol polarization, where Pol is the VV, HH or VH polarization channel. For the polarization time-series images, _DA under different scatterer mechanisms ω can be obtained by the following formula:

in

In the formula, SM represents the scattering mechanism, N represents the number of SAR images, and the upper horizontal line represents the average value, where the complex vector ω of the scattering mechanism is the characteristic vector u _i obtained above;

Use the BEST algorithm at the pixel level for fully polarized PS targets:

为该点相位质量最好的像元；In the formula

is the pixel with the best phase quality at this point;

Use the BEST algorithm for dual polarized PS targets:

Step 2, γ criterion optimization of DS objective:

The coherence γ has a corresponding functional relationship with the standard deviation of the interferometric phase, which is the evaluation standard of the interferometric phase quality. The DS target is optimized by using the γ standard measurement. The larger the γ value, the better the phase quality;

The γ of the interferogram for Pol polarization is calculated by the following formula:

In the formula, S ₁ and S ₂ represent the two SAR images that constitute the interferogram, E represents the expectation operator, and the mathematical expression of multi-polarization γ is:

Polarization optimization using the gamma criterion for the DS target to calculate the average coherence

In the formula, M represents the number of interferograms, and j represents the jth interferogram;

Use the BEST algorithm at the pixel level for fully polarized DS targets:

为该点相位质量最好的像元；In the formula

is the pixel with the best phase quality at this point;

Use the BEST algorithm at the pixel level for dual-polarization DS targets:

The phase values of the PS and DS pixels with the best phase quality in each pixel are combined to form the final polarization-optimized interferometric phase.

The beneficial effect of the present invention is that: firstly, the present invention distinguishes PS and DS targets through homogeneous pixel identification, uses eigenvalue decomposition and BEST algorithm to search for optimal scattering mechanism, and applies different strategies to optimize PS and DS targets. By making full use of the different responses of different scatterers to polarization signals and using multi-polarization information, the phase quality of the interferogram is significantly improved, the phase noise of the interferogram is reduced, and the interference fringes are clearer. The number of deformation monitoring points has been significantly increased, and through the optimization algorithm, monitoring points can be obtained in places that were not detected by the original image.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present invention. Those skilled in the art can also obtain other drawings based on these drawings without creative work.

Fig. 1 is a flow chart of a polarization timing InSAR method based on coherent matrix adaptive decomposition provided by an embodiment of the present invention

Fig. 2 is a comparison diagram of differential interference details before and after polarization optimization provided by an embodiment of the present invention;

Fig. 3 is a comparison diagram of TPC values before and after polarization optimization provided by an embodiment of the present invention;

Fig. 4 is the TPC statistical curve before and after optimization provided by the embodiment of the present invention;

Fig. 5 is a comparison diagram of time-series settlement monitoring results before and after polarization optimization provided by the embodiment of the present invention;

FIG. 6 is a detailed comparison diagram of timing settlement results before and after polarization optimization provided by an embodiment of the present invention.

Detailed ways

The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

As shown in Figure 1, a polarization time-series InSAR method based on coherent matrix adaptive decomposition, this embodiment uses a total of 34 dual-polarization Sentinels covering Shanghai Pudong Airport from May 15, 2016 to December 24, 2017 -1 data (33 scene interferograms obtained by using single main image method), the number of image pixels is 700×2400.

The specific implementation plan is as follows:

Step 1: Acquire dual-polarization Sentinel-1 data, preprocess Sentinel-1 SAR images and divide PS and DS targets.

The original VV and VH polarimetric SAR images are subjected to preprocessing steps such as registration, cropping, and geocoding, and a statistical test algorithm is used to identify homogeneous pixels and divide PS and DS targets.

Step two:

According to the PS and DS targets obtained in step 1, the polarization scattering matrix S is constructed:

In the formula, S represents the scattering matrix, S _vv represents the single-view complex SAR image under VV polarization, and S _vh represents the single-view complex SAR image under VH polarization.

The polarized scattering vector _ki is composed under the Pauli basis:

k _i =[S _{vv, i} , 2S _{vh, i} ] ^T

In the formula, S _vv,i represents the single-view complex SAR image under the i-th VV polarization, S _vh,i represents the single-view complex SAR image under the i-th VH polarization, and T represents the matrix transposition operation.

According to the polarization scattering vector, the polarization coherence matrix of the PS target and the polarization vector interferometric phase of the PS target are obtained:

where k _j represents the jth polarized scattering vector.

where H represents the conjugate transpose, and k _j represents the jth polarized scattering vector.

I=μ ₁ ·μ ₂ ^*

where * is the complex conjugate operator, and I is the polarization vector interferometric phase.

Obtain the DS polarization coherence matrix and perform MMSE filtering:

In the formula, k ₁ and k ₂ represent the scattering vectors under the Pauli basis, T ₁₁ and T ₂₂ represent the coherence matrix, and Ω ₁₂ represents the polarization coherence matrix. In the MMSE polarization filtering of homogeneous pixels, it is first necessary to calculate the span image of the polarization coherence matrix T:

span=Trace(T ₁₁ )+Trace(T ₂₂ )

In the formula, Trace represents the trace of the matrix, and the filter weight coefficient b is calculated by the value of span as follows:

In the formula, SHPS represents a homogeneous pixel, and the coef value is 0.51. Finally, the MMSE filtering of the homogeneous pixel is completed by the following formula, and the MMSE filtering is completed for the polarization coherence matrix T ^DS of the DS target:

T ^DS ＝T _mean +b(T _ori -T _mean )

The eigenvalue decomposition is performed after combining the PS and DS polarization coherence matrices:

表示平均时间极化相干矩阵，λ_i表示特征值，u_i表示特征向量，num表示极化通道个数。Sentinel-1数据为双极化数据，num＝2，会得到两个特征值λ₁≥λ₂≥0和对应的特征向量u₁、u₂。In the formula

Represents the average time polarization coherence matrix, λ _i represents the eigenvalue, u _i represents the eigenvector, and num represents the number of polarization channels. Sentinel-1 data is dual-polarization data, num=2, two eigenvalues λ ₁ ≥λ ₂ ≥0 and corresponding eigenvectors u ₁ , u ₂ will be obtained.

Perform D _A criterion optimization on the PS objective:

Calculate the original VV and VH polarimetric SAR images respectively, and the eigenvectors u ₁ and u ₂ obtained above are used as the _DA value of the PS target in the complex vector ω of the scattering mechanism.

表示Pol极化下幅度的标准差，

表示Pol极化下幅度的均值，其中Pol可以是VV和VH极化通道。对于极化时序影像，不同散射体机制ω下的D_A可通过下式得到：In the formula

Indicates the standard deviation of the amplitude under the Pol polarization,

Indicates the mean value of the amplitude under Pol polarization, where Pol can be both VV and VH polarized channels. For polarization time-series images, D _A under different scatterer mechanisms ω can be obtained by the following formula:

in

In the formula, SM represents the scattering mechanism, N represents the number of SAR images, and the upper horizontal line represents the average value, where the complex vector ω of the scattering mechanism is the characteristic vector u _i obtained above.

Use the BEST algorithm at the pixel level for fully polarized PS targets:

为该点相位质量最好的像元。In the formula

The pixel with the best phase quality at this point.

Perform gamma criterion optimization on the DS objective:

The γ of the interferogram for VV or VH polarization can be calculated by the following formula:

In the formula, S ₁ and S ₂ represent the two SAR images that constitute the interferogram, and E represents the expectation operator, which is realized by taking the average of homogeneous pixels in the present invention. The mathematical expression of multipolar γ is:

Polarization optimization using the gamma criterion for the DS target to calculate the average coherence

In the formula, M represents the number of interferograms, and j represents the jth interferogram.

Use the BEST algorithm at the pixel level for fully polarized DS targets:

为该点相位质量最好的像元。In the formula

The pixel with the best phase quality at this point.

The phase values of the PS and DS pixels with the best phase quality in each pixel are combined to form the final polarization-optimized interferometric phase.

Step 3: Monitoring the surface deformation based on the polarization-optimized differential interferogram;

Use the polarization-optimized interferometric phase of eigenvalue decomposition to calculate the time-series average coherence, select high-quality pixels, use the polarization-optimized differential interferogram to perform polarization time-series interferometry, obtain deformation monitoring points, and complete the alignment Deformation monitoring in the study area. In order to reflect the monitoring effect of the present invention, it is compared with the monitoring effect of the commonly used PS-InSAR method (VV polarization data based on Sentinel-1). In this comparative example, the same parameters are set for the time series processing of the differential interferogram.

First, compare the interferogram results. The original VV polarization interferogram, as shown in Figure 2, has a large level of interference phase noise and poor quality. There are interference fringes in the edge area of the station building, but they are not smooth and continuous. There are basically no interference fringes in the area of distributed scatterers, and there is a serious decoherence phenomenon. After the polarization optimization is carried out by the method proposed by the present invention, the fringes of the interference pattern become clearer and more continuous, the edge contour details of buildings and other structures are preserved, and the interference phase noise is significantly reduced, and the spatial distribution is continuous, smooth and stable.

Then, the temporal phase coherence (TPC) is used to quantitatively evaluate the quality of the interferogram, and the phase quality of the differential interferogram is calculated and evaluated by using the noise level of the pixel in all interferograms, which can be calculated by the following formula:

In the formula, M is the number of interferograms, and ψ _noisei is the noise phase corresponding to the i-th interferogram. The window for ψ _noisei (considering the SAR image resolution and the surface coverage of the study area, this study takes a 15×15 window) is estimated based on the field pixel interferometric phase value.

As shown in Figure 3 and Figure 4, the larger the TPC value, the better the phase quality of the interferogram. It can be clearly observed that the method proposed in the present invention will obtain more high-quality pixels, which further proves that the method proposed in the present invention can effectively improve the phase quality of the differential interference atlas, and increase the time required for subsequent monitoring point selection. The number of high-quality cells.

In this embodiment, TPC>0.9 is selected as the threshold for high-quality pixel point selection, and the commonly used open source software StaMPS is used for timing analysis and deformation calculation, and finally the deformation monitoring results are shown in Figure 5. It can be seen from the figure that the trend of the deformation results obtained by the inversion of the two methods is basically the same. The monitoring points of the two monitoring methods are 20845 and 58697 respectively. Compared with the original VV polarization, the number of monitoring points obtained by the method of the present invention has increased by 182%. , the method proposed in the present invention has obtained more details of the surface deformation, and some details of the monitoring results of the surface deformation are shown in Figure 6. Take the two local maps in Figure 5 for comparison, and set them as local 1 and local 2 for comparison, C and D It can be seen that the density of monitoring points has increased significantly in the region.

The method of the invention has the characteristics of obvious optimization effect, high density of monitoring points, fast calculation efficiency, etc., and is consistent with the result trend of the common single polarization monitoring method, which illustrates the effectiveness and reliability of the method of the invention.

Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention also intends to include these modifications and variations.

Claims (7)

1. A polarization time sequence InSAR method based on coherent matrix self-adaptive decomposition is characterized by comprising the following steps:

s1, acquiring and researching a multi-polarization time sequence SAR image in a monitoring area, preprocessing the SAR image and identifying homogeneous pixels to obtain PS and DS targets;

s2, carrying out polarization optimization under different criteria according to PS and DS targets, and comprising the following steps:

s21, constructing a polarization scattering vector for the multi-polarization data under the Pauli base;

s22, constructing a PS target polarization coherent matrix T ^PS And the polarization vector interference phase of the PS target;

s23, constructing a polarization coherent matrix T for the DS target ^DS Carrying out polarization self-adaptive MMSE filtering;

s24, decomposing a polarization coherent matrix eigenvalue;

s25, aiming at PS and DS targets, respectively adopting D _A Carrying out polarization optimization according to the criterion and the gamma criterion;

and S3, calculating time sequence coherence by using the polarization-optimized interference phase decomposed by the characteristic value, selecting high-quality pixels, and performing polarization time sequence interference measurement by using a polarization-optimized differential interference pattern to obtain deformation monitoring points so as to complete deformation monitoring on a research area.

2. The polarization time sequence InSAR method based on the coherent matrix adaptive decomposition as claimed in claim 1, wherein the specific method of step S1 is as follows:

after acquiring and researching a multi-polarization time sequence SAR image of a monitoring area, carrying out registration, cutting and geocoding pretreatment on the SAR image, dividing the multi-polarization SAR image into PS and DS targets under respective polarization by using the intensity information of the multi-polarization time sequence SAR image and a homogeneous pixel identification method based on statistical test, and then using amplitude dispersion D _A Criterion for D in DS target _A And converting the pixels with the value less than 0.33 into PS targets, and acquiring a union set of identification results of different polarization homogeneous points after carrying out homogeneous pixel identification on the multi-polarization SAR image because different ground objects have different intensity information in different polarizations, so as to obtain the final PS and DS targets.

3. The polarization time sequence InSAR method based on the coherent matrix adaptive decomposition as claimed in claim 1, wherein the specific method of step S21 is as follows:

the polarized radar transmits and receives electromagnetic pulse signals in a vertical mode and a horizontal mode, records the electromagnetic pulse signals as different polarized images according to the mode of transmitting and receiving signals during antenna imaging, has four polarization modes of VV, VH, HH and HV in total, and constructs a scattering matrix S according to polarization data:

in which s represents a scattering matrix, s _hh Representing a monoscopic complex SAR image under HH polarization, S _vv Representing single-view complex SAR images under VV polarization, S _vh Representing single-view complex SAR images, S, under VH polarization _hv Representing a single-view complex SAR image under HV polarization;

for the fully polarized data, a Pauli-based scattering vector k is constructed _q The following were used:

wherein q represents the total polarization, according to the reciprocity theorem S _vh ＝S _hv Thus using only S _vh Represents cross polarization, and T represents matrix transposition operation;

for dual polarization data, there are a number of different polarization combining methods:

for the SAR image data combined by HH and VV, constructing a scattering vector k under Pauli base _d The following were used:

constructing a Pauli-based scattering vector k for dual-polarized SAR image data combining co-polarization (HH, VV) and cross-polarization (VH, HV) _d The following were used:

k _d ＝[S _xx ，2S _vh ] ^T

wherein d represents a dual polarization; s _xx Representing a single-view complex SAR image under HH or VV polarization;

the polarization vector interference is different from the single polarization scalar interference, the single polarization scalar interference is conjugate multiplication between a first image and a second image which are registered, the polarization vector interference needs to construct respective scattering vectors of the two polarization images, a scattering machine is introduced to make an outer product to obtain an interference phase, and a scattering mechanism is represented by a unit complex projection vector omega;

for fully polarized data, ω _q Is obtained by the following formula:

in the formula, alpha, beta and psi respectively represent polarization parameters of scattering mechanism vectors, and j and e respectively represent imaginary number units and natural constants;

for dual polarized data, ω _d Is obtained by the following formula:

for polarization vector interference, unit complex vector omega of two image specific scattering mechanisms is used _i For the scattering vectors k of the first and second images _i Respectively projecting by the following formula to obtain mu _i ：

μ _i ＝ω _i ^H ·k _i ，i＝1，2

Wherein i represents the ith image, H represents the conjugate transpose, and μ _i Components representing respective specific scattering mechanisms, similar to a single-view complex image in a single polarization.

4. The polarization time sequence InSAR method based on coherent matrix adaptive decomposition according to claim 3, wherein the specific method of step S22 is as follows:

adopting a time coherent matrix, and directly constructing a polarization coherent matrix for a PS target:

in the formula k _j Represents the jth polarization scattering vector;

the polarization vector interference phase is generated on the basis of two complex scalars:

I＝μ ₁ ·μ ₂ ^*

in the formula, the complex conjugate operator is shown.

5. The polarization time sequence InSAR method based on the coherent matrix adaptive decomposition as claimed in claim 4, wherein the specific method of step S23 is as follows:

construction of a polarized coherent matrix T for DS targets ^DS

In the formula k ₁ 、k ₂ Denotes the scattering vector under Pauli basis, T ₁₁ And T ₂₂ Denotes the coherence matrix, Ω ₁₂ Representing a polarization coherent matrix, in MMSE polarization filtering of homogeneous pixels, firstly calculating a span image of the polarization coherent matrix T:

span＝Trace(T ₁₁ )+Trace(T ₂₂ )

wherein, trace represents the Trace of the matrix, and the filtering weight coefficient b is calculated according to the following formula by the value of span:

SHPS represents homogeneous pixel, coef value is 0.51, MMSE filtering of the homogeneous pixel is completed through the following formula, and a polarization coherent matrix T of a DS target is subjected to ^DS And (3) completing MMSE filtering:

T ^DS ＝T _mean +b(T _ori -T _mean )

in the formula T _mean Representing the average coherence matrix, T, of homogeneous pixels within a window _ori Representing the original coherence matrix.

6. The polarization time sequence InSAR method based on the coherent matrix adaptive decomposition as claimed in claim 5, wherein the specific method of step S24 is as follows:

will T ^PS And T ^DS Combining the polarized coherent matrixes of the two targets to form a polarized coherent matrix T of the whole image for eigenvalue decomposition, and decomposing the time sequence average polarized coherent matrix into:

in the formula

Representing the mean time-polarized coherence matrix, λ _i Representing a characteristic value, u _i Representing a characteristic vector, and num representing the number of polarization channels;

when fully polarized data is used, num =3, resulting in three eigenvalues λ ₁ ≥λ ₂ ≥λ ₃ Not less than 0 and corresponding feature vector u ₁ 、u ₂ 、u ₃ ；

When dual polarization data is used, num =2, two eigenvalues λ are obtained ₁ ≥λ ₂ Not less than 0 and corresponding feature vector u ₁ 、u ₂ 。

7. The polarization time sequence InSAR method based on coherent matrix adaptive decomposition as claimed in claim 6, wherein the specific method of step S24 is as follows:

step one, D of PS target _A And (3) criterion optimization:

obtaining a phase quality evaluation index D by taking the ratio of the standard deviation and the mean value of the time sequence SAR image amplitude _A For optimizing PS targets, D _A The smaller the value, the better the phase quality, which is given by:

in the formula

Represents the standard deviation of the amplitude at Pol polarization,

means for the amplitude at Pol polarization, where Pol isVV, HH or VH polarization channels, D under different scatterer mechanisms omega for polarization time sequential images _A Obtained by the following formula:

wherein

In the formula, SM represents a scattering mechanism, N represents the number of SAR images, an upper transverse line represents an average value, wherein a scattering mechanism complex vector omega is the obtained characteristic vector u _i ；

Using the BEST algorithm at the pixel level for a fully polarized PS target:

in the formula

The pixel with the best phase quality at the point is obtained;

the BEST algorithm is used for the dual polarization PS target:

step two, optimizing the gamma criterion of the DS target:

the coherence gamma and the interference phase standard deviation have a corresponding functional relation and are evaluation standards of interference phase quality, a DS target is optimized by gamma standard measurement, and the phase quality is better when the gamma value is larger;

γ for Pol polarized interferograms is calculated by:

in the formula S ₁ And S ₂ Representing the two SAR images constituting the interferogram, E representing the desired operator, the mathematical expression of the multipolarization γ is:

polarization optimization for DS targets using gamma criterion, computing average coherence

Wherein M represents the number of interferograms, and j represents the jth interferogram;

the BEST algorithm is used at the pixel level for the fully-polarized DS target:

in the formula

The pixel with the best phase quality at the point is obtained;

the BEST algorithm is used at the pixel level for the dual-polarized DS target:

and combining the phase values of the PS and DS pixels with the best phase quality under each pixel element to form the final polarization optimization interference phase.

Images