Multi-Channel SAR-GMTI Method Robust to Coregistration Error of SAR Images

Robust differential interferometric synthetic aperture radar deformation phase estimation method based on covariance matrix information of multiple pixels

“…1 as an example, we assume that there is no coregistration error in the azimuth direction, the number of the neighbouring pixels used to construct the joint complex interferometric phase data vector is 3 and the processing pixel in the interferograms is marked as (i, j). The joint complex D-InSAR phase data vector jp(i, j) similar to the data vector constructed in [11,16,17], can be formulated as…”

“…It should be noted that in [11,16,17] the data vector is directly formulated by the SAR complex image data, whereas the joint complex D-InSAR phase data vector jp(i, j) in our approach is formulated by complex interferograms. Therefore for D-InSAR processing the SAR images need to be coarsely coregistered firstly.…”

“…Hence, for the D-InSAR processing, the differential interferometric phase vector method is also robust to coregistration errors, including SAR image coregistration errors and/or interferograms coregistration errors. The detailed explanation of robustness can be found in [11].…”

“…Since the directions of coregistration errors are random and hard to estimate, it is difficult to identify the large eigenvalues and small eigenvalues of the eigenspectrum of C jp (i, j) clearly. From (11), the minimum norm of the joint covariance matrix C jp (i, j) can be used to estimate the deformation phase…”

“…Eigenanalysis of the covariance matrix is widely used in signal processing. For example, it was exploited for SAR ground moving target indication (GMTI) [11][12][13]. Signal subspace processing is also used to equalise the SAR images blindly [14].…”