CN103454636B - Differential interferometric phase estimation method based on multi-pixel covariance matrixes - Google Patents
Differential interferometric phase estimation method based on multi-pixel covariance matrixes Download PDFInfo
- Publication number
- CN103454636B CN103454636B CN201310404990.0A CN201310404990A CN103454636B CN 103454636 B CN103454636 B CN 103454636B CN 201310404990 A CN201310404990 A CN 201310404990A CN 103454636 B CN103454636 B CN 103454636B
- Authority
- CN
- China
- Prior art keywords
- deformation
- interferometric phase
- differential interferometry
- phase
- interferometric
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention discloses a differential interferometric phase estimation method based on multi-pixel covariance matrixes. The differential interferometric phase estimation method mainly solves the problem that an existing differential interferometric phase generating algorithm is not stable in registration errors. The differential interferometric phase estimation method includes the first step of inputting images and parameters, the second step of registering interferometric phase diagrams, the third step of obtaining the interferometric phase diagrams through secondary interference, the fourth step of obtaining image windows, the fifth step of constructing repeated interferometric phase joint data vectors, the sixth step of estimating the covariance matrixes, the seventh step of estimating correlation coefficient matrixes, the eighth step of carrying out matrix characteristic decomposition, the ninth step of constructing cost functions, and the tenth step of estimating differential interferometric phases. The method has the advantages of restoring pixel information in a self-adaptation mode under the condition that the interferometric phase diagrams have the registration errors, accurately estimating the differential interferometric phases, and being capable of effectively reducing the influence of the registration errors on the differential interferometric phases.
Description
Technical field
The invention belongs to communication technical field, further relate to the differential interferometry phase estimation method based on many pixels covariance matrix in the radar exploration technique field.Under the present invention may be used for the condition that in differential interferometry synthetic-aperture radar (Differential Interferometric Synthetic Aperture Radar, D-InSAR), interferometric phase image registration accuracy is very poor, sane estimation difference interferometric phase.
Background technology
D-InSAR is the very fast remote sensing technology of development in recent years, there is the advantage of round-the-clock, round-the-clock, wide area high resolution detection Ground Deformation, relative to existing GPS (Global Position System, and measurement of the level means GPS), there are huge advantage and wide application prospect, have been widely used in the fields such as city settlement monitoring, earthquake deformation measurement, volcanic explosion.
Image registration accuracy determines the order of accuarcy of the differential interferometry phase place that D-InSAR obtains, and differential interferometry phase place has directly reacted the deformation quantity on ground, and the registration accuracy of interferometric phase image is the key factor affecting the generation of differential interferometry phase place.Existing differential interferometry phase estimation method is mainly based on conventional InSAR treatment scheme, within conventional InSAR processing requirements registration accuracy reaches 1/10 pixel, and need complicated Interpolation Process, when registration error is more than 1/10 pixel, inaccurate differential interferometry phase place will be obtained.
M.Massonnet, the people such as R.M.Goldstein and H.A.Zebker are at article " Mapping small elevation changes over large areas:differential interferometry " (Journal of Geophysical Research, 1989, vol.94:9183-9191) the middle method of estimation proposing a kind of differential interferometry phase place based on existing InSAR interferometric phase generation.The method first carries out smart registration operation to SAR image and interferometric phase image, then does second order difference to the interferometric phase image obtained, and finally obtains differential interferometry phase place to obtained result filtering.The weak point that the method exists is: when interferometric phase image has larger registration error to exist, result accurately cannot be obtained, and need when obtaining differential interferometry phase place to carry out filtering, registration and filtering are carried out step by step, registration error can lose differential interferometry phase estimation precision, and phase-filtering step can not recover the loss of significance that registration error is brought, its process errors is constantly accumulation.
Liao Mingsheng, Lu Lijun, Wang Yan and Li Deren proposes a kind of method estimating difference interferometric phase based on coherent point target analysis in article " research based on the InSAR technology for detection earth's surface miniature deformation that point target is analyzed " (" urban geology ", 2006,1 (2): 38-41).The weak point that the method exists is: only have interferometric phase image before and after deformation all smart registration time, just can obtain differential interferometry phase place accurately, and for the interferometric phase image before and after the deformation of rough registration, the noise that the differential interferometry phase diagram that the method obtains contains is larger.
Patented claim " the earth's surface three-dimensional deformation monitoring method based on InSAR and gps data the merge " (date of application: on 02 05th, 2010 that He'nan University proposes, application number: 201010106794, publication number: CN101770027A) in disclose a kind of in conjunction with gps data merge differential interferometry phase place preparation method.The method revises SAR satellite orbital error by the coordinate transformation relation setting up GPS and SAR data, then obtains geography information by the interferometric phase image of unwrapping accuracy registration, finally merges InSAR and GPS deformation data and obtains differential interferometry phase place.The weak point that the method exists is: when revising SAR satellite orbital error with gps data, the necessary registration completely of corresponding pixel, when fusion InSAR and GPS deformation data, and the necessary registration completely of pixel that same landform is corresponding.
Summary of the invention
When the present invention is directed to the generation differential interferometry phase place of above-mentioned prior art existence, the defect of the necessary accuracy registration of the interferometric phase image before and after deformation, proposes a kind of differential interferometry phase estimation method based on many pixels covariance matrix.The present invention builds covariance matrix by the information consolidation of multiple pixel, again signal subspace is obtained to estimated covariance matrix feature decomposition, finally utilize the method for signal subspace fitting to estimate difference interferometric phase, ensure that interferometric phase image before and after deformation is under the very poor condition of registration accuracy, the robust iterative of differential interferometry phase place, completes differential interferometry phase filtering simultaneously.
For achieving the above object, key step of the present invention is as follows:
(1) input picture and parameter:
The level land interferometric phase image that goes before target area deformation 1a) differential interferometry synthetic-aperture radar obtained respectively and after deformation is input to system;
1b) systematic parameter of differential interferometry synthetic-aperture radar is input to system;
(2) interferometric phase image registration:
2a) deduct rear data line with the every data line in the interferometric phase image of level land that goes before deformation respectively, obtain removing level land row differential interferometry phase diagram before deformation;
2b) deduct a rear column data with each column data in the row differential interferometry phase diagram of level land that goes before deformation respectively, obtain the interferometric phase gradient map before deformation;
2c) deduct rear data line with the every data line in the interferometric phase image of level land that goes after deformation respectively, obtain removing level land row differential interferometry phase diagram after deformation;
2d) deduct a rear column data with each column data in the row differential interferometry phase diagram of level land that goes after deformation respectively, obtain the interferometric phase gradient map after deformation;
2e) with the interferometric phase gradient map before deformation for reference picture, registration process is carried out to the interferometric phase gradient map after deformation, the side-play amount when interferometric phase gradient map before the deformation obtaining registration and after deformation and registration;
2f) by all pixels of going in the interferometric phase image of level land after deformation, a side-play amount during translation registration, to obtain before the deformation of rough registration interferometric phase image after interferometric phase image and deformation;
(3) interferometric phase image that secondary is interfered is obtained:
3a) be multiplied by scale factor by the data of pixels all in interferometric phase image before the deformation of rough registration, before obtaining the deformation of rough registration, become mark interferometric phase image;
3b) by the data of all pixels in interferometric phase image after the deformation of rough registration, become the data of corresponding pixel points in mark interferometric phase image before deducting the deformation of rough registration, obtain the interferometric phase image that secondary is interfered;
(4) image window is obtained:
In the interferometric phase image that secondary is interfered, centered by pixel to be estimated, regular length is block radius, obtains a square window;
(5) multiple interferometric phase associating data vector is built:
5a) in square window, choose arbitrarily a pixel, by selected pixel and around it data of neighbor pixel form a line, obtain the phase vectors of selected pixel;
5b) by each element in the phase vectors of selected pixel, the index that to ask with natural Exponents e after being multiplied with imaginary unit be the end, obtains the index vector of selected pixel;
5c) before each element of the index vector of selected pixel, insert positive integer 1, obtain the multiple interferometric phase associating data vector of pixel to be estimated;
5d) travel through all pixels in square window, judge whether the multiple interferometric phase associating data vector obtaining all pixels, if so, then perform step (6), otherwise, perform step 5a);
(6) following formula is adopted, estimate covariance matrix:
Wherein, C represents estimated covariance matrix, and l (i, j) represents the multiple interferometric phase associating data vector of the i-th row jth column element in square window, i=1,2 ..., M, j=1,2,, M, M represent the radius of square window, and H represents and does conjugate transposition operation;
(7) coefficient of coherence matrix is estimated:
All elements value in covariance matrix is taken absolute value, obtains the coefficient of coherence matrix of pixel to be estimated;
(8) feature decomposition operation is carried out to coefficient of coherence matrix, obtain eigenwert and the proper vector of coefficient of coherence matrix;
(9) following formula is adopted, structure cost function:
Wherein, J represents constructed cost function,
represent differential interferometry phase place
corresponding differential interferometry phase vectors, N represents the number of large eigenwert in the eigenwert of coefficient of coherence matrix, β
prepresent large eigenwert characteristic of correspondence vector in the proper vector of coefficient of coherence matrix, p=1,2 ..., N, β
qrepresent large eigenwert characteristic of correspondence vector in the proper vector of coefficient of coherence matrix, q=1,2 ..., N, H represent and do conjugate transposition operation,
represent that Hadamard Hadamard amasss;
(10) difference interferometric phase is estimated:
(-π, π] scope in, search differential interferometry phase place
, find the maximal value of cost function, by the differential interferometry phase place corresponding to cost function maximal value
as the differential interferometry phase place of pixel to be estimated.
The present invention has the following advantages compared with prior art:
First, the cost function that the present invention utilizes the proper vector of covariance matrix to construct is to estimate difference interferometric phase, what overcome prior art estimates in difference interferometric phase method based on coherent point target analysis, in interferometric phase image rough registration situation, the inaccurate defect of differential interferometry phase place obtained, the present invention is made to have good robustness to interferometric phase image registration error, have when interferometric phase image existence very large registration error, still can obtain the advantage of differential interferometry phase place accurately.
Second, the present invention utilizes many pixels joint mapping covariance matrix, overcome in the differential interferometry phase estimation method based on the generation of InSAR interferometric phase of prior art, the defect that the differential interferometry phase loss that registration error causes cannot be recovered by phase filtering, make the present invention can adaptive recovery differential interferometry phase information, having phase information utilizes fully, differential interferometry phase estimation advantage accurately.
Accompanying drawing explanation
Fig. 1 is process flow diagram of the present invention;
Fig. 2 be adopt prior art InSAR interference technique emulation generate remove level land interferometric phase image;
Fig. 3 is the differential interferometry phase diagram of the Generation of simulating data adopting the present invention and prior art process interferometric phase image accuracy registration;
Fig. 4 adopts the present invention and prior art process interferometric phase image registration error to be the differential interferometry phase diagram of the Generation of simulating data of 1 pixel;
Fig. 5 is the differential interferometry phase diagram adopting the present invention and prior art process measured data to generate.
Embodiment
Below in conjunction with accompanying drawing, the present invention will be further described.
With reference to accompanying drawing 1, specific embodiment of the invention step is as follows:
Step 1, input picture and parameter.
By differential interferometry synthetic-aperture radar obtain target area deformation before with deformation after remove level land interferometric phase image, the imaging number of times to target area and the base length of differential interferometry synthetic-aperture radar are input to system.Go base length corresponding to level land interferometric phase image to be less than scarcely to go the base length that level land interferometric phase image is corresponding before than deformation after deformation.
Step 2, interferometric phase image registration.
The registration of interferometric phase image, is carry out registration to the gradient map of two width interferometric phase images, then carries out translation according to the side-play amount of gradient map to interferometric phase image.In the present invention, only need the registration accuracy of two width interferometric phase images to reach a pixel, significantly reduce the difficulty of image registration.
The concrete steps of interferometric phase image registration are as follows:
The first step, deducts rear data line with the every data line in the interferometric phase image of level land that goes before deformation, obtains removing level land row differential interferometry phase diagram before deformation;
Second step, deducts a rear column data with each column data in the row differential interferometry phase diagram of level land that goes before deformation, obtains the interferometric phase gradient map Q before deformation;
3rd step, deducts rear data line with the every data line in the interferometric phase image of level land that goes after deformation, obtains removing level land row differential interferometry phase diagram after deformation;
4th step, deducts a rear column data with each column data in the row differential interferometry phase diagram of level land that goes after deformation, obtains the interferometric phase gradient map S after deformation;
5th step, utilizes following formula, obtains the cross-correlation matrix of the interferometric phase gradient map before deformation and the interferometric phase gradient map after deformation:
L=IFFT2(FFT2(Q)×conj(FFT2(S)))
Wherein, L represents the cross-correlation matrix of the interferometric phase gradient map after the interferometric phase gradient map before deformation and deformation, IFFT2 () represents that doing two-dimentional inverse Fourier transform operates, FFT2 () represents that doing two-dimensional Fourier transform operates, Q represents the interferometric phase gradient map data matrix before deformation, S represents the interferometric phase gradient map data matrix after deformation, and conjugate operation is got in conj () expression;
6th step, maximizing from all elements of cross-correlation matrix L, obtains the coordinate figure that in cross-correlation matrix L, maximal value element is corresponding;
7th step, with the coordinate figure that maximal value element in cross-correlation matrix L is corresponding, deducts the coordinate figure that the central element of cross-correlation matrix L is corresponding, obtains side-play amount during registration;
8th step, according to the side-play amount of the interferometric phase gradient map after deformation during registration, does integral translation to the level land interferometric phase image that goes after deformation, obtain before the deformation of rough registration go level land interferometric phase image and deformation after remove level land interferometric phase image.
Step 3, obtains the interferometric phase image that secondary is interfered.
Dual-interferometry is done to the interferometric phase image of two width registrations, with the interferometric phase image after deformation for reference, interferometric phase image before deformation is adjusted in ratio identical with it, and scale factor and radar to the imaging number of times of target area and the base length of radar relevant.
The concrete steps obtaining the interferometric phase image that secondary is interfered are as follows:
The first step, when the imaging number of times of difference interference synthetic aperture radar to target area equals 2, scale factor equals 1;
Second step, when the imaging number of times of difference interference synthetic aperture radar to target area is greater than 2, according to the following formula, calculates scale factor:
Wherein, k represents scale factor, the base length of the differential interferometry synthetic-aperture radar of going level land interferometric phase image corresponding after B represents deformation, the base length of the differential interferometry synthetic-aperture radar of going level land interferometric phase image corresponding before A represents deformation;
3rd step, by the data of removing all pixels in the interferometric phase image of level land before deformation, is multiplied with scale factor k, obtains the change mark before deformation and removes level land interferometric phase image;
4th step, by the data of removing all pixels in the interferometric phase image of level land after deformation, deducts the data that the change mark before deformation goes corresponding pixel points in the interferometric phase image of level land, obtains the interferometric phase image that secondary is interfered.
Step 4, obtains image window.
In the interferometric phase image that secondary is interfered, centered by pixel to be estimated, with regular length 3 ~ 7 pixels for block radius, obtain a square window.The square window that size is 9 × 9 is have chosen in example of the present invention.
Step 5, builds multiple interferometric phase associating data vector.
Use current pixel point and pixel around thereof when building multiple interferometric phase associating data vector, make use of the information of current pixel point fully, thus can the information of adaptive reply current pixel point when making to generate differential interferometry phase place.
The concrete steps building multiple interferometric phase associating data vector are as follows:
5a) in square window, choose arbitrarily a pixel, by selected pixel and around it data of 1 ~ 8 neighbor pixel form a line, obtain the phase vectors of selected pixel, in this invention, 8 neighbor pixels that have employed surrounding build jointly;
5b) by each element in the phase vectors of selected pixel, the index that to ask with natural Exponents e after being multiplied with imaginary unit be the end, obtains the index vector of selected pixel;
5c) before each element of the index vector of selected pixel, insert positive integer 1, obtain the multiple interferometric phase associating data vector of pixel to be estimated;
5d) travel through all pixels in square window, judge whether the multiple interferometric phase associating data vector of all pixels obtains, if all obtained, then performs step 6, otherwise, perform step 5a).
Step 6, estimate covariance matrix.
Under normal circumstances, directly cannot obtain covariance matrix, so utilize the space sample of covariance matrix on average to replace statistical average.Following formula estimate covariance matrix is adopted in example of the present invention:
Wherein, C represents estimated covariance matrix, l (i, j) the multiple interferometric phase associating data vector of the i-th row jth column element in square window is represented, i=1,2,, M, j=1,2,, M, M represent the radius of square window, in this invention, M=4, H represent and do conjugate transposition operation.
Step 7, estimates coefficient of coherence matrix.
Due in interferometric phase image, noise power is much smaller than signal power, and very little diagonal angle loading affects very little on the eigenwert of matrix and proper vector, so utilize covariance matrix to estimate coefficient of coherence matrix.All elements value in covariance matrix C is taken absolute value, obtains the coefficient of coherence matrix R of pixel to be estimated.
Step 8, carries out feature decomposition operation to coefficient of coherence matrix R, obtains eigenwert and the proper vector of coefficient of coherence matrix R.
Step 9, estimates difference interferometric phase.
In this invention, use the method for signal subspace fitting to estimate difference interferometric phase, first construct cost function J by the base vector of signal subspace.
Adopt following formula, structure cost function:
Wherein, J represents constructed cost function, and N represents the number of large eigenwert in the eigenwert of coefficient of coherence matrix, N=9, β in this invention
prepresent large eigenwert characteristic of correspondence vector in the proper vector of coefficient of coherence matrix, p=1,2 ..., 9, β
qrepresent large eigenwert characteristic of correspondence vector in the proper vector of coefficient of coherence matrix, q=1,2 ..., 9, H represents and does conjugate transposition operation,
represent that Hadamard Hadamard amasss,
represent differential interferometry phase place
corresponding differential interferometry phase vectors, the concrete steps building differential interferometry phase vectors are as follows:
The first step, by the differential interferometry phase place of pixel to be estimated
, the index that to ask with natural Exponents e after being multiplied with imaginary unit j be the end, obtains the exponential quantity of pixel to be estimated
;
Second step, by the exponential quantity of positive integer 1 with pixel to be estimated
order forms a line, and obtains the single differential interferometry phase vectors of pixel to be estimated;
3rd step, chooses 9 single differential interferometry phase vectors, is formed a line by all elements in 9 single differential interferometry vectors, obtains the differential interferometry phase vectors of pixel to be estimated
.
Step 10, estimates difference interferometric phase.
(-π, π] scope in, search differential interferometry phase place
, find the maximal value of cost function J, by the differential interferometry phase place corresponding to cost function J maximal value
as the differential interferometry phase place of pixel to be estimated.
Below in conjunction with emulation experiment, effect of the present invention is described further.
1, simulated conditions:
With reference to accompanying drawing 1, by differential interferometry synthetic-aperture radar, simulation imaging is carried out to areal and obtain three width SAR image, after adopting the InSAR interference treatment of prior art, obtain before deformation respectively and remove level land interferometric phase image after deformation, it can be used as the input picture of system.
Fig. 2 be adopt prior art InSAR interference technique emulation generate remove level land interferometric phase image, wherein, Fig. 2 (a) for adopt prior art InSAR interference technique emulation generate deformation before remove level land interferometric phase image, Fig. 2 (b) be adopt prior art InSAR interference technique emulation emulation generation deformation after remove level land interferometric phase image.Because maximum landform shape becomes 0.1m, thus in Fig. 2 (a) and Fig. 2 (b) to go level land interferometric phase image to be similar to identical, simulation parameter arranges as follows:
| Systematic parameter | Parameter value |
| Wavelength | 0.03m |
| Antenna length | 2m |
| Pulse repetition rate | 500Hz |
| Platform speed | 80m/s |
| Radar downwards angle of visibility | 45° |
| Pulse width | 5us |
| Platform flying height | 6000m |
| Base length | 0.8m |
| Range resolution | 1m |
| Largest deformation amount | 0.1m |
2, analysis of simulation result:
Fig. 3 is the differential interferometry phase diagram of the Generation of simulating data adopting the present invention and prior art process interferometric phase image accuracy registration, wherein, Fig. 3 (a) is for adopting the differential interferometry phase diagram of the Generation of simulating data of prior art process interferometric phase image accuracy registration, abscissa axis represent distance to, unit is pixel, axis of ordinates represent orientation to, unit is pixel, in the image in left side, the gray-scale value of pixel represents the size of differential interferometry phase place, the size of differential interferometry phase fluctuation in figure, reflect the number of differential interferometry phase noise, the image bar on right side represents the corresponding relation of gray-scale value in differential interferometry phase value and left-side images, the size of the numeric representation differential interferometry phase place of image right, Fig. 3 (b) is for adopting the differential interferometry phase diagram of the Generation of simulating data of process interferometric phase image accuracy registration of the present invention.
Contrast accompanying drawing 3(a) and accompanying drawing 3(b), large than Fig. 3 (b) of the fluctuation of the differential interferometry phase place that deformation region is corresponding in Fig. 3 (a), i.e. large than Fig. 3 (b) of the differential interferometry phase noise of generation in Fig. 3 (a).Thus can obtain, when interferometric phase image accuracy registration, the differential interferometry phase diagram quality that the present invention obtains, higher than prior art, has self-adaptive recovery differential interferometry phase information, differential interferometry phase estimation advantage accurately.
Fig. 4 adopts the present invention and prior art process interferometric phase image registration error to be the differential interferometry phase diagram of the Generation of simulating data of 1 pixel, wherein, Fig. 4 (a) is the differential interferometry phase diagram of the Generation of simulating data of 1 pixel for adopting prior art process interferometric phase image registration error, the differential interferometry phase diagram of Fig. 4 (b) be employing process interferometric phase image registration error of the present invention be Generation of simulating data of 1 pixel.
Contrast accompanying drawing 4(a) and accompanying drawing 4(b), be all made up of noise in Fig. 4 (a), do not comprise any useful information, and Fig. 4 (b) still can obtain differential interferometry phase diagram clearly, and differential interferometry phase fringes is more clear.Thus can obtain, when there is the error of 1 pixel when between interferometric phase image, the present invention can obtain the good differential interferometry phase diagram of quality, is significantly better than prior art.Contrast accompanying drawing 3(b) and accompanying drawing 4(b), in the differential interferometry phase diagram of Fig. 4 (b), differential interferometry phase fluctuation is only less times greater than Fig. 3 (b), and can obtain thus, the present invention has good robustness to interferometric phase image registration error.
Below in conjunction with measured data result, the present invention is described further.
Process measured data used, be European Space Agency ENVISAT satellite respectively on February 23rd, 2009 and May 4 to the SAR imaging results in centralItaly L ' Aquila area, and the digital elevation diagram data that US National Aeronautics and Space Administration (NASA) publishes.
Fig. 5 is the differential interferometry phase diagram adopting the present invention and prior art process measured data to generate, wherein, the differential interferometry phase diagram that Fig. 5 (a) generates for adopting prior art process measured data, Fig. 5 (b) is the differential interferometry phase diagram adopting process measured data of the present invention to generate.
Contrast accompanying drawing 5(a) and accompanying drawing 5(b), only to see fuzzy differential interferometry phase fringes in the differential interferometry phase diagram of Fig. 5 (a), and containing a lot of phase noises in figure, differential interferometry phase fringes is clearly had in the interferometric phase image of Fig. 5 (b), differential interferometry phase place is comparatively level and smooth, only containing a small amount of interferometric phase noise, recover the partial information of differential interferometry phase place.Can obtain thus, the present invention is when processing interferometric phase image registration error the unknown, still the differential interferometry phase diagram than prior art better quality can be obtained, complete the operation of differential interferometry phase filtering simultaneously, describing the present invention can adaptive recovery differential interferometry phase information, further demonstrate the robustness of the present invention to registration error, and describe practicality of the present invention and validity simultaneously.
Claims (8)
1., based on the differential interferometry phase estimation method of many pixels covariance matrix, comprise the steps:
(1) input picture and parameter:
The level land interferometric phase image that goes before target area deformation 1a) differential interferometry synthetic-aperture radar obtained respectively and after deformation is input to system;
1b) systematic parameter of differential interferometry synthetic-aperture radar is input to system;
(2) interferometric phase image registration:
2a) deduct rear data line with the every data line in the interferometric phase image of level land that goes before deformation respectively, obtain removing level land row differential interferometry phase diagram before deformation;
2b) deduct a rear column data with each column data in the row differential interferometry phase diagram of level land that goes before deformation respectively, obtain the interferometric phase gradient map before deformation;
2c) deduct rear data line with the every data line in the interferometric phase image of level land that goes after deformation respectively, obtain removing level land row differential interferometry phase diagram after deformation;
2d) deduct a rear column data with each column data in the row differential interferometry phase diagram of level land that goes after deformation respectively, obtain the interferometric phase gradient map after deformation;
2e) with the interferometric phase gradient map before deformation for reference picture, registration process is carried out to the interferometric phase gradient map after deformation, the side-play amount when interferometric phase gradient map before the deformation obtaining registration and after deformation and registration;
2f) by all pixels of going in the interferometric phase image of level land after deformation, a side-play amount during translation registration, to obtain before the deformation of rough registration interferometric phase image after interferometric phase image and deformation;
(3) interferometric phase image that secondary is interfered is obtained:
3a) be multiplied by scale factor by the data of pixels all in interferometric phase image before the deformation of rough registration, before obtaining the deformation of rough registration, become mark interferometric phase image;
3b) by the data of all pixels in interferometric phase image after the deformation of rough registration, become the data of corresponding pixel points in mark interferometric phase image before deducting the deformation of rough registration, obtain the interferometric phase image that secondary is interfered;
(4) image window is obtained:
In the interferometric phase image that secondary is interfered, centered by pixel to be estimated, regular length is block radius, obtains a square window;
(5) multiple interferometric phase associating data vector is built:
5a) in square window, choose arbitrarily a pixel, by selected pixel and around it data of neighbor pixel form a line, obtain the phase vectors of selected pixel;
5b) by each element in the phase vectors of selected pixel, the index that to ask with natural Exponents e after being multiplied with imaginary unit be the end, obtains the index vector of selected pixel;
5c) before each element of the index vector of selected pixel, insert positive integer 1, obtain the multiple interferometric phase associating data vector of pixel to be estimated;
5d) travel through all pixels in square window, judge whether the multiple interferometric phase associating data vector obtaining all pixels, if so, then perform step (6), otherwise, perform step 5a);
(6) following formula is adopted, estimate covariance matrix:
Wherein, C represents estimated covariance matrix, and l (i, j) represents the multiple interferometric phase associating data vector of the i-th row jth column element in square window, i=1,2 ..., M, j=1,2,, M, M represent the radius of square window, and H represents and does conjugate transposition operation;
(7) coefficient of coherence matrix is estimated:
All elements value in covariance matrix is taken absolute value, obtains the coefficient of coherence matrix of pixel to be estimated;
(8) feature decomposition operation is carried out to coefficient of coherence matrix, obtain eigenwert and the proper vector of coefficient of coherence matrix;
(9) following formula is adopted, structure cost function:
Wherein, J represents constructed cost function,
represent differential interferometry phase place
corresponding differential interferometry phase vectors, N represents the number of large eigenwert in the eigenwert of coefficient of coherence matrix, β
prepresent large eigenwert characteristic of correspondence vector in the proper vector of coefficient of coherence matrix, p=1,2 ..., N, β
qrepresent large eigenwert characteristic of correspondence vector in the proper vector of coefficient of coherence matrix, q=1,2 ..., N, H represent and do conjugate transposition operation,
represent that Hadamard Hadamard amasss;
(10) difference interferometric phase is estimated:
(-π, π] scope in, search differential interferometry phase place
find the maximal value of cost function, by the differential interferometry phase place corresponding to cost function maximal value
as the differential interferometry phase place of pixel to be estimated.
2. the differential interferometry phase estimation method based on many pixels covariance matrix according to claim 1, is characterized in that: step 1b) described in the systematic parameter of differential interferometry synthetic-aperture radar comprise imaging number of times and base length; Described imaging number of times refers to the number of times of differential interferometry synthetic-aperture radar to target area imaging, base length refer to respectively with the length of the differential interferometry synthetic-aperture radar baseline going level land interferometric phase image corresponding before deformation and after deformation.
3. the differential interferometry phase estimation method based on many pixels covariance matrix according to claim 1, is characterized in that: step 2e) described in the concrete steps of interferometric phase gradient map registration as follows:
The first step, utilizes following formula, obtains the cross-correlation matrix of the interferometric phase gradient map before deformation and the interferometric phase gradient map after deformation:
L＝IFFT2(FFT2(R)×conj(FFT2(S)))
Wherein, L represents the cross-correlation matrix of the interferometric phase gradient map after the interferometric phase gradient map before deformation and deformation, IFFT2 () represents that doing two-dimentional inverse Fourier transform operates, FFT2 () represents that doing two-dimensional Fourier transform operates, R represents the interferometric phase gradient map data matrix before deformation, S represents the interferometric phase gradient map data matrix after deformation, and conjugate operation is got in conj () expression;
Second step, maximizing from all elements of cross-correlation matrix L, obtains the coordinate figure that in cross-correlation matrix L, maximal value element is corresponding;
3rd step, with the coordinate figure that maximal value element in cross-correlation matrix L is corresponding, deducts the coordinate figure that the central element of cross-correlation matrix L is corresponding, obtains side-play amount during registration;
4th step, by the interferometric phase gradient map after deformation, the side-play amount in units of pixel during a translation registration, the registration of the interferometric phase gradient map before completing deformation and after deformation.
4. the differential interferometry phase estimation method based on many pixels covariance matrix according to claim 1, is characterized in that: step 3a) described in the computing method of scale factor as follows:
The first step, when the imaging number of times of difference interference synthetic aperture radar to target area equals 2, scale factor equals 1;
Second step, when the imaging number of times of difference interference synthetic aperture radar to target area is greater than 2, according to the following formula, calculates scale factor:
Wherein, k' represents scale factor, the base length of the differential interferometry synthetic-aperture radar of going level land interferometric phase image corresponding after B represents deformation, the base length of the differential interferometry synthetic-aperture radar of going level land interferometric phase image corresponding before A represents deformation.
5. the differential interferometry phase estimation method based on many pixels covariance matrix according to claim 1, is characterized in that: regular length described in step (4) is 3 ~ 7 pixels.
6. the differential interferometry phase estimation method based on many pixels covariance matrix according to claim 1, is characterized in that: step 5a) in be used for building the neighbor pixel of multiple interferometric phase associating data vector number be 1 ~ 8.
7. the differential interferometry phase estimation method based on many pixels covariance matrix according to claim 1, is characterized in that: described in step (9), the preparation method of the differential interferometry phase vectors that differential interferometry phase place is corresponding is as follows:
The first step, by the differential interferometry phase place of pixel to be estimated, the index that to ask with natural Exponents e after being multiplied with imaginary unit be the end, obtains the exponential quantity of pixel to be estimated;
Second step, forms a line the exponential quantity order of positive integer 1 with pixel to be estimated, obtains the single differential interferometry phase vectors of pixel to be estimated;
3rd step, choose and build the single differential interferometry phase vectors using the number of pixel identical when multiple interferometric phase combines data vector, all elements in selected single differential interferometry vector is formed a line, obtains the differential interferometry phase vectors of pixel to be estimated.
8. the differential interferometry phase estimation method based on many pixels covariance matrix according to claim 1, is characterized in that: the number of large eigenwert described in step (9) is identical with the number that the multiple interferometric phase of structure combines data vector all pixels point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310404990.0A CN103454636B (en) | 2013-09-08 | 2013-09-08 | Differential interferometric phase estimation method based on multi-pixel covariance matrixes |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310404990.0A CN103454636B (en) | 2013-09-08 | 2013-09-08 | Differential interferometric phase estimation method based on multi-pixel covariance matrixes |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103454636A CN103454636A (en) | 2013-12-18 |
| CN103454636B true CN103454636B (en) | 2015-05-20 |
Family
ID=49737211
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310404990.0A Active CN103454636B (en) | 2013-09-08 | 2013-09-08 | Differential interferometric phase estimation method based on multi-pixel covariance matrixes |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103454636B (en) |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103996175B (en) * | 2014-05-13 | 2017-02-15 | 西安电子科技大学 | Forest or urban area high-resolution interference phase filtering method |
| CN105469368A (en) * | 2015-11-30 | 2016-04-06 | 中国人民解放军国防科学技术大学 | Interferometric phase filtering method |
| CN105974411B (en) * | 2016-04-28 | 2018-03-09 | 电子科技大学 | High-voltage power transmission tower top tilt displacement monitoring method based on DINSAR |
| CN108663678B (en) * | 2018-01-29 | 2022-01-18 | 西北农林科技大学 | Multi-baseline InSAR phase unwrapping algorithm based on mixed integer optimization model |
| CN108303735A (en) * | 2018-01-30 | 2018-07-20 | 单新建 | The earthquake deformation acquisition methods of subband interferometry based on optimized parameter setting |
| CN108627832A (en) * | 2018-05-11 | 2018-10-09 | 电子科技大学 | A method of passway for transmitting electricity Ground Deformation is extracted based on multiple timings SAR image |
| CN109541596B (en) * | 2018-11-28 | 2022-05-20 | 中国电子科技集团公司电子科学研究院 | InSAR image processing method and device based on deep learning algorithm |
| CN111595293A (en) * | 2020-05-29 | 2020-08-28 | 山东交通学院 | Surface deformation monitoring method and system based on multi-source monitoring data fusion |
| CN118549934B (en) * | 2024-07-30 | 2024-10-11 | 云南大学 | Phase estimation optimization method, device and equipment |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101551455B (en) * | 2009-05-13 | 2011-10-19 | 西安电子科技大学 | 3D terrain imaging system of interferometric synthetic aperture radar and elevation mapping method thereof |
| CN101770027B (en) * | 2010-02-05 | 2012-05-16 | 河海大学 | Ground surface three-dimensional deformation monitoring method based on InSAR and GPS data fusion |
-
2013
- 2013-09-08 CN CN201310404990.0A patent/CN103454636B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN103454636A (en) | 2013-12-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103454636B (en) | Differential interferometric phase estimation method based on multi-pixel covariance matrixes | |
| CN110058237B (en) | InSAR point cloud fusion and three-dimensional deformation monitoring method for high-resolution SAR image | |
| CN104123464B (en) | Method for inversion of ground feature high elevation and number of land subsidence through high resolution InSAR timing sequence analysis | |
| CN103439708B (en) | Polarized InSAR interferogram estimation method based on generalized scattering vector | |
| Goel et al. | A distributed scatterer interferometry approach for precision monitoring of known surface deformation phenomena | |
| CN103885059B (en) | A kind of multi-baseline interference synthetic aperture radar three-dimensional rebuilding method | |
| CN106772342A (en) | A kind of Timing Difference radar interference method suitable for big gradient surface subsidence monitoring | |
| CN109212522B (en) | High-precision DEM inversion method and device based on double-base satellite-borne SAR | |
| CN109031301A (en) | Alpine terrain deformation extracting method based on PSInSAR technology | |
| CN104007439B (en) | Interferential circular SAR elevation estimation processing method | |
| CN115201825B (en) | Atmospheric delay correction method in InSAR (interferometric synthetic aperture radar) inter-seismic deformation monitoring | |
| CN109239710B (en) | Method and device for acquiring radar elevation information and computer-readable storage medium | |
| CN102680972A (en) | Method and device for monitoring surface deformation and data processing equipment | |
| CN103018741A (en) | Interferometric synthetic aperture radar (InSAR) imaging and flat ground removing integral method based on back projection | |
| CN111650579A (en) | InSAR mining area three-dimensional deformation estimation method and device for rock migration parameter adaptive acquisition and medium | |
| CN110109105A (en) | A method of the InSAR technical monitoring Ground Deformation based on timing | |
| CN103323840A (en) | Method for time alignment between interference SAR echo data and platform motion and gesture data | |
| CN103308914B (en) | One-station fixed bistatic interference synthetic aperture radar (SAR) processing method | |
| CN114858238A (en) | Underground water storage and transformation monitoring and early warning system | |
| CN113281749A (en) | Time sequence InSAR high-coherence point selection method considering homogeneity | |
| CN107144213A (en) | The big magnitude three-D sequential deformation method of estimation in mining area and device based on SAR intensity images | |
| Li et al. | An adaptive phase optimization algorithm for distributed scatterer phase history retrieval | |
| CN103809180A (en) | Azimuth pre-filtering processing method for Interferometric Synthetic Aperture Radar (InSAR) topographic survey | |
| Ji et al. | Applying InSAR and GNSS data to obtain 3-D surface deformations based on iterated almost unbiased estimation and Laplacian smoothness constraint | |
| CN116068511B (en) | Deep learning-based InSAR large-scale system error correction method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |