InSAR (Interferometric Synthetic Aperture Radar) sequence time decoherence suppression method based on complex covariance matrix

Abstract

The invention belongs to the technical field of synthetic aperture radars, and relates to an InSAR sequence time decoherence suppression method based on a complex covariance matrix, and the specific process is as follows: for a plurality of input SLC SAR image complex vectors, extracting a single-pixel p-dimensional complex scattering vector at a pixel x, and generating an initial complex covariance matrix; the method comprises the following steps: generating a rectangular search window Wx by taking a pixel x as a center, aiming at multiple input SLC SAR image complex vectors, extracting a p-dimensional complex scattering vector of a single pixel at a pixel y in the search window, and generating a reference complex covariance matrix; enabling the pixel y to traverse each pixel except the pixel x in the rectangular search window Wx, and calculating a maximum likelihood test statistic Q (x, y) according to the initial complex covariance matrix and the reference complex covariance matrix; and based on a complex covariance matrix # imgabs0 # estimated by Q (x, y), extracting the intensity, the phase and the correlation coefficient of the InSAR sequence after de-coherence suppression at the pixel x from the complex covariance matrix # imgabs0 #, and finally obtaining a time sequence de-coherence suppression result of the complete SAR interferogram sequence.

Description

Technical Field

[0001] The invention belongs to the technical field of synthetic aperture radar, and in particular relates to an InSAR sequence temporal decorrelation suppression method based on a complex covariance matrix. Background Art

[0002] Synthetic Aperture Radar (SAR) is a high-resolution imaging radar. Its all-weather, all-day operation and wide-area coverage make it an indispensable technology for Earth observation. Synthetic Aperture Radar Interferometry (InSAR), developed from SAR technology, constructs an interferogram containing phase difference information by acquiring two SAR images of the same area with different spatial or temporal baselines. By accurately calculating the interferometric phase information, InSAR can infer the three-dimensional spatial coordinates and deformation characteristics of the observed target.

[0003] In the process of acquiring InSAR images, if a long time baseline is used, it is very easy to cause significant temporal decoherence. That is, due to the change of surface characteristics over time, the scattering characteristics of the radar echo signal during the two SAR imaging are inconsistent, resulting in a significant decrease in the correlation coefficient of the generated interferogram, affecting the imaging quality and measurement accuracy.

[0004] Currently, widely used methods for suppressing temporal decoherence utilize techniques such as permanent scatterers (PS) or distributed scatterers (DS). These techniques select areas with stable surface features for processing to improve the coherence of the interferogram. However, the PS method relies on stable scatterers, resulting in insufficient spatial density and poor performance, particularly in non-urban areas. While the DS method offers wider coverage, it requires high scatterer stability and suffers from algorithmic complexity. Furthermore, filtering methods are often used to reduce the phase noise generated by temporal decoherence. Current filtering methods often process the interferogram in the spatial domain and fail to utilize temporal information, resulting in limited effectiveness in suppressing temporal decoherence. Some improved multi-temporal filtering methods also process only the intensity of the interferogram and fail to preserve its phase information, rendering them unhelpful for subsequent InSAR processing steps. Summary of the Invention

[0005] In order to solve the problem of temporal decorrelation suppression of the above-mentioned SAR interferogram sequence, the present invention proposes a temporal decorrelation suppression method for InSAR sequences based on the complex covariance matrix. This method can determine homogeneous pixels in the spatiotemporal joint domain according to the statistical characteristics of the complex covariance matrix of the scattering unit, and estimate the complex covariance matrix after suppressing temporal decorrelation based on the maximum likelihood criterion, thereby deriving more accurate interferogram intensity, phase and correlation coefficient information.

[0006] The technical solutions for implementing the present invention are as follows:

[0007] In a first aspect, the present invention provides a method for suppressing temporal decorrelation of InSAR sequences based on a complex covariance matrix, the specific process of which is as follows:

[0008] For the input multiple SLC SAR image complex vectors, the following processing is performed at each pixel x:

[0009] Take out the p-dimensional complex scattering vector z of a single pixel at pixel x j (x), and generate the initial complex covariance matrix C(x), where j = 1, 2, ..., p is the complex value of the j-th image at pixel x;

[0010] Generate a rectangular search window W centered at pixel x x , for the input multiple SLC SAR image complex vectors, extract the single pixel p-dimensional complex scattering vector z at pixel y in the search window j (y), generate the reference complex covariance matrix C(y);

[0011] Let pixel y traverse the rectangular search window W x For each pixel except pixel x, the maximum likelihood test statistic Q(x,y) is calculated based on the initial complex covariance matrix C(x) and the reference complex covariance matrix C(y);

[0012] The complex covariance matrix of the estimate Q(x,y) is measured based on the maximum likelihood test Then the intensity, phase and correlation coefficient of the InSAR sequence after decorrelation suppression at pixel x are extracted;

[0013] Finally, the temporal decorrelation suppression results of the complete SAR interferogram sequence are obtained.

[0014] Optionally, the maximum likelihood test statistic Q(x,y) of the present invention is:

[0015]

[0016] Here, |·| represents the determinant of the matrix, and n represents the size of the complex correlation calculation window.

[0017] Optionally, the complex covariance matrix estimated by the maximum likelihood test measurement Q(x,y) of the present invention is The specific process is:

[0018] Calculate the weight w(x,y) based on the maximum likelihood test metric Q(x,y);

[0019] Based on the weight of each pixel in the window, the complex covariance matrix corresponding to pixel x is estimated

[0020] Optionally, the weight w(x,y) of the present invention is:

[0021] w(x,y)=-ln Q(x,y)

[0022] The complex covariance matrix for:

[0023]

[0024] in, is the normalization coefficient.

[0025] Optionally, the present invention uses the complex covariance matrix Extract the intensity, phase and correlation coefficient of the InSAR sequence after decoherence suppression at pixel x,

[0026]

[0027] in, represents the intensity of the i-th image at pixel x, represents the intensity of the j-th image at pixel x, represents the phase of the interference pattern generated by the i-th image and the j-th image at pixel x, Correlation coefficient of the interference pattern generated by the i-th image and the j-th image at pixel x.

[0028] In a second aspect, the present invention provides an InSAR sequence temporal decorrelation suppression device based on a complex covariance matrix, comprising: a complex covariance matrix generation module, a likelihood statistics calculation module, and a decorrelation suppression module;

[0029] The complex covariance matrix generation module is used to extract the p-dimensional complex scattering vector z of a single pixel at pixel x based on the input multiple SLC SAR image complex vectors j (x), and generate the initial complex covariance matrix C(x), where j = 1, 2, ..., p is the complex value of the j-th image at pixel x; generate a rectangular search window W centered on pixel x x , for the input multiple SLC SAR image complex vectors, extract the single pixel p-dimensional complex scattering vector z at pixel y in the search window j (y), generate the reference complex covariance matrix C(y);

[0030] A likelihood statistic calculation module is used to calculate the maximum likelihood test statistic Q(x,y) based on the initial complex covariance matrix C(x) and the reference complex covariance matrix C(y);

[0031] Decorrelation suppression module, based on the maximum likelihood test measurement Q (x, y) estimated complex covariance matrix The intensity, phase and correlation coefficient of the InSAR sequence after decoherence suppression at pixel x are extracted; finally, the temporal decoherence suppression result of the complete SAR interferogram sequence is obtained.

[0032] Beneficial effects:

[0033] The temporal decorrelation suppression method mentioned in the present invention can determine homogeneous pixels in the spatiotemporal joint domain based on the statistical characteristics of the complex covariance matrix of the scattering unit, making full use of the time domain information of the original SAR sequence; and based on the maximum likelihood criterion, the complex covariance matrix after suppressing temporal decorrelation is estimated, thereby retaining the phase information of the interferogram sequence. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0035] Figure 1 It is a flow chart of the steps of the present invention.

[0036] Figure 2 is the unprocessed original InSAR interferometric phase image.

[0037] Figure 3 This is the InSAR interferometric phase image obtained after processing using the method of the present invention. DETAILED DESCRIPTION

[0038] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0039] It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments may be combined with each other; and, based on the embodiments in this disclosure, all other embodiments obtained by persons of ordinary skill in the art without creative work are within the scope of protection of this disclosure.

[0040] It should be noted that various aspects of the embodiments within the scope of the appended claims are described below. It should be apparent that the aspects described herein can be embodied in a wide variety of forms, and any specific structure and / or function described herein is merely illustrative. Based on this disclosure, it should be understood by those skilled in the art that an aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects described herein can be used to implement an apparatus and / or practice a method. In addition, other structures and / or functionalities other than one or more of the aspects described herein can be used to implement this apparatus and / or practice this method.

[0041] In the embodiment of the present application, the user input is a complex vector S generated by p SLC SAR images = [S1, S2, ..., S p ], where S1, S2, …, S p is a single SLC SAR image complex matrix with dimension l row ×l col , l row and l col The complex covariance matrix represents the number of rows and columns of the image, the size of the complex correlation calculation window is n, and the side length of the search window is t; the output is the intensity, phase and correlation coefficient of the SAR interferometer pair after decoherence suppression. The embodiment of the present application is a method for suppressing temporal decoherence of InSAR sequences based on the complex covariance matrix, and the specific steps are as follows:

[0042] For the input multiple SLC SAR image complex vectors, the following processing is performed at each pixel x:

[0043] Take out the p-dimensional complex scattering vector z of a single pixel at pixel x j (x), and generate the initial complex covariance matrix C(x), where j = 1, 2, ..., p is the complex value of the j-th image at pixel x;

[0044] Generate a rectangular search window W centered at pixel x x , for the input multiple SLC SAR image complex vectors, extract the single pixel p-dimensional complex scattering vector z at pixel y in the search window j (y), generate the reference complex covariance matrix C(y);

[0045] Let pixel y traverse the rectangular search window W x For each pixel except pixel x, the maximum likelihood test statistic Q(x,y) is calculated based on the initial complex covariance matrix C(x) and the reference complex covariance matrix C(y);

[0046] The complex covariance matrix of the estimate Q(x,y) is measured based on the maximum likelihood test The intensity, phase and correlation coefficient of the InSAR sequence after decoherence suppression at pixel x are extracted;

[0047] Finally, the temporal decorrelation suppression results of the complete SAR interferogram sequence are obtained.

[0048] The above process is described in detail below. Figure 1 As shown:

[0049] Step 1: Get the p-dimensional complex scattering vector Z(x) of a single pixel

[0050] At pixel x, from the input l row ×l col Extract the p-dimensional complex scattering vector Z(x) of a single pixel from the p-dimensional SLC SAR image complex vector

[0051] Z(x)=[z1(x),z2(x),…,z p (x)] T (1)

[0052] Among them, z j (x)(j=1,2,…,p)=S j (x), is the complex value of the j-th image at pixel x.

[0053] Step 2: Generate an initial complex covariance matrix based on the scattering vector Z(x)

[0054] Calculate the window size n based on the input complex correlation and generate the initial complex covariance matrix C(x)

[0055]

[0056] Step 3: Generate a rectangular search window

[0057] With pixel x as the center, generate a rectangular search window W of size t×t x .

[0058] Step 4: Take the p-dimensional complex scattering vector Z(y) of a single pixel in the search window.

[0059] In the search window W x Take a pixel y and follow step 1 to get the complex scattering vector Z(y)

[0060] Z(y)=[z1(y),z2(y),…,z p (y)] T (3)

[0061] Among them, z j (y)(j=1,2,…,p)=S j(y) is the complex value of the j-th image at pixel y.

[0062] Step 5: Generate the reference complex covariance matrix

[0063] Generate the reference complex covariance matrix C(y) at pixel y

[0064]

[0065] Step 6: Calculate the maximum likelihood test statistic

[0066] Calculate the maximum likelihood test statistic Q(x,y) based on the initial complex covariance matrix C(x) and the reference complex covariance matrix C(y)

[0067]

[0068] Here, |·| represents the determinant of the matrix.

[0069] The derivation process of the maximum likelihood test statistic is as follows:

[0070] The complex covariance matrix C(x) obeys the complex Wishart distribution, and the probability density function is as follows:

[0071]

[0072] in, Σ(x)=E{Z(x)Z(x) H}, is the covariance matrix of pixel x.

[0073] Based on the maximum likelihood derivation of the probability density function, there are two assumptions:

[0074] Assumption H0: The covariance matrices of the two distributions are the same Σ(x)=Σ(y)=Σ

[0075] Assumption H1: The covariance matrices of the two distributions are different Σ(x)≠Σ(y)

[0076] Under the assumption H0, the joint likelihood function is:

[0077] f(C(x),C(y)|n,Σ)=w(C(x)|n,Σ)·w(C(y)|n,Σ) (7)

[0078] The log-likelihood is:

[0079]

[0080] It can be obtained that under the assumption H0, the maximum likelihood estimate is:

[0081]

[0082] Similarly, under the assumption H1, the maximum estimate is:

[0083]

[0084] From this we can get the maximum likelihood ratio:

[0085]

[0086] Taking the maximum likelihood ratio as the maximum likelihood test statistic, we get:

[0087]

[0088] Step 7: Calculate the weight w(x,y) based on the maximum likelihood test statistic Q(x,y)

[0089] w(x,y)=-ln Q(x,y) (13)

[0090] Since the larger the value of the maximum likelihood test statistic, the higher the similarity between the two pixels, the weight is defined as the complex logarithm of the maximum likelihood test statistic so as to quantitatively perform weighted averaging according to the degree of similarity.

[0091] Step 8: Weighted average to achieve decoherence suppression

[0092] Traverse steps 4 to 7 in the generated search window, calculate the weight pixel by pixel, and perform weighted averaging to achieve decorrelation suppression to obtain the estimated complex covariance matrix The formula is as follows

[0093]

[0094] in, is the normalization coefficient.

[0095] Step 9: Extract the intensity, phase and correlation coefficient of the InSAR sequence after decoherence suppression

[0096] From the complex covariance matrix Extract the intensity, phase and correlation coefficient after decoherence suppression at pixel x

[0097]

[0098] in, represents the intensity of the i-th image at pixel x, represents the intensity of the j-th image at pixel x, represents the phase of the interference pattern generated by the i-th image and the j-th image at pixel x, Correlation coefficient of the interference pattern generated by the i-th image and the j-th image at pixel x.

[0099] Step 10, traverse all l row ×l col pixels, repeat steps 1 to 9 to obtain the temporal decorrelation suppression result of the complete SAR interferogram sequence.

[0100] In a second aspect, the present invention provides an InSAR sequence temporal decorrelation suppression device based on a complex covariance matrix, comprising: a complex covariance matrix generation module, a likelihood statistics calculation module, and a decorrelation suppression module;

[0101] The complex covariance matrix generation module is used to extract the p-dimensional complex scattering vector z of a single pixel at pixel x based on the input multiple SLC SAR image complex vectors j (x), and generate the initial complex covariance matrix C(x), where j = 1, 2, ..., p is the complex value of the j-th image at pixel x; generate a rectangular search window W centered on pixel x x , for the input multiple SLC SAR image complex vectors, extract the single pixel p-dimensional complex scattering vector z at pixel y in the search window j (y), generate the reference complex covariance matrix C(y);

[0102] A likelihood statistic calculation module is used to calculate the maximum likelihood test statistic Q(x,y) based on the initial complex covariance matrix C(x) and the reference complex covariance matrix C(y);

[0103] Decorrelation suppression module, based on the maximum likelihood test measurement Q (x, y) estimated complex covariance matrix The intensity, phase and correlation coefficient of the InSAR sequence after decoherence suppression at pixel x are extracted; finally, the temporal decoherence suppression result of the complete SAR interferogram sequence is obtained.

[0104] Next, an implementation example is given with specific parameters.

[0105] In this example, the input data consists of three TerraSAR-X images, capturing the Beijing Regent Hotel and some surrounding buildings. The correlation calculation window size is set to n = 9, and the search window side length is set to t = 9.

[0106] First, the initial complex covariance matrix is ​​generated by steps 1 and 2. Then, the search window is generated by step 3. Then, the reference complex covariance matrix is ​​generated by steps 4 and 5. Next, the maximum likelihood test statistic and weights are calculated by steps 6 and 7. Finally, the temporal decorrelation suppression of the SAR interferometric sequence is completed by steps 8, 9, and 10.

[0107] Figure 2Given the original interference phase of two interference pairs in the interferogram sequence, Figure 3 The interference phase after processing by the method proposed in this invention is shown. By comparison, it can be seen that the phase noise of the two groups of interference pairs has been significantly reduced, effectively suppressing the temporal decoherence of the original interference pattern.

[0108] In summary, the above are only preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for suppressing temporal decorrelation of InSAR sequences based on complex covariance matrix, characterized in that: The specific process is: For the input multiple SLC SAR image complex vectors, the following processing is performed at each pixel x: Take out the p-dimensional complex scattering vector z of a single pixel at pixel x j (x), and generate the initial complex covariance matrix C(x), where j = 1, 2, ..., p is the complex value of the j-th image at pixel x; Generate a rectangular search window W centered at pixel x x , for the input multiple SLC SAR image complex vectors, extract the single pixel p-dimensional complex scattering vector z at pixel y in the search window j (y), generate the reference complex covariance matrix C(y); Let pixel y traverse the rectangular search window W x For each pixel except pixel x, the maximum likelihood test statistic Q(x,y) is calculated based on the initial complex covariance matrix C(x) and the reference complex covariance matrix C(y); The complex covariance matrix of the estimate Q(x,y) is measured based on the maximum likelihood test The intensity, phase and correlation coefficient of the InSAR sequence after decoherence suppression at pixel x are extracted; Finally, the temporal decorrelation suppression results of the complete SAR interferogram sequence are obtained.

2. The InSAR sequence temporal decorrelation suppression method based on the complex covariance matrix according to claim 1, characterized in that: The maximum likelihood test statistic Q(x,y) is: Here, |·| represents the determinant of the matrix, and n represents the size of the complex correlation calculation window.

3. The InSAR sequence temporal decorrelation suppression method based on the complex covariance matrix according to claim 2, characterized in that: The complex covariance matrix estimated based on the maximum likelihood test measurement Q(x,y) The specific process is: Calculate the weight w(x,y) based on the maximum likelihood test metric Q(x,y); Based on the weight of each pixel in the window, the complex covariance matrix corresponding to pixel x is estimated 4. The InSAR sequence temporal decorrelation suppression method based on the complex covariance matrix according to claim 3, characterized in that: The weight w(x,y) is: w(x,y)=-lnQ(x,y) The complex covariance matrix for: in, is the normalization coefficient.

5. The InSAR sequence temporal decorrelation suppression method based on the complex covariance matrix according to claim 4, characterized in that: From the complex covariance matrix Extract the intensity, phase and correlation coefficient of the InSAR sequence after decoherence suppression at pixel x in, represents the intensity of the i-th image at pixel x, represents the intensity of the j-th image at pixel x, represents the phase of the interference pattern generated by the i-th image and the j-th image at pixel x, Correlation coefficient of the interference pattern generated by the i-th image and the j-th image at pixel x.

6. A device for suppressing temporal decorrelation of InSAR sequences based on a complex covariance matrix, characterized in that: include: Complex covariance matrix generation module, likelihood statistics calculation module and decorrelation suppression module; The complex covariance matrix generation module is used to extract the p-dimensional complex scattering vector z of a single pixel at pixel x based on the input multiple SLC SAR image complex vectors j (x), and generate the initial complex covariance matrix C(x), where j = 1, 2, ..., p is the complex value of the j-th image at pixel x; generate a rectangular search window W centered on pixel x x , for the input multiple SLC SAR image complex vectors, extract the single pixel p-dimensional complex scattering vector z at pixel y in the search window j (y), generate the reference complex covariance matrix C(y); A likelihood statistic calculation module is used to calculate the maximum likelihood test statistic Q(x,y) based on the initial complex covariance matrix C(x) and the reference complex covariance matrix C(y); Decorrelation suppression module, based on the maximum likelihood test measurement Q (x, y) estimated complex covariance matrix The intensity, phase and correlation coefficient of the InSAR sequence after decoherence suppression at pixel x are extracted; finally, the temporal decoherence suppression result of the complete SAR interferogram sequence is obtained.

7. The InSAR sequence temporal decorrelation suppression device based on complex covariance matrix according to claim 6, characterized in that: The maximum likelihood test statistic Q(x,y) is: Here, |·| represents the determinant of the matrix, and n represents the size of the complex correlation calculation window.

8. The InSAR sequence temporal decorrelation suppression device based on complex covariance matrix according to claim 7, characterized in that: The complex covariance matrix estimated based on the maximum likelihood test measurement Q(x,y) The specific process is: Calculate the weight w(x,y) based on the maximum likelihood test metric Q(x,y); Based on the weight of each pixel in the window, the complex covariance matrix corresponding to pixel x is estimated 9. The InSAR sequence temporal decorrelation suppression device based on complex covariance matrix according to claim 8, characterized in that: The weight w(x,y) is: w(x,y)=-lnQ(x,y) The complex covariance matrix for: in, is the normalization coefficient.

10. The InSAR sequence temporal decorrelation suppression device based on complex covariance matrix according to claim 9, characterized in that: From the complex covariance matrix Extract the intensity, phase and correlation coefficient of the InSAR sequence after decoherence suppression at pixel x, in, represents the intensity of the i-th image at pixel x, represents the intensity of the j-th image at pixel x, represents the phase of the interference pattern generated by the i-th image and the j-th image at pixel x, Correlation coefficient of the interference pattern generated by the i-th image and the j-th image at pixel x.

Images