Image Auto Coregistration and Interferogram Estimation Using Matrix Fitting

In this paper, we proposed a novel method based on matrix fitting to estimate the synthetic aperture radar interferometry (InSAR) or synthetic aperture sonar interferometry (InSAS) interferometric phase. In this method, we formulate generalized steering vector at first. And then, taking advantage of the coherence information of neighboring pixel pairs, covariance matrix of complex data vector is estimated. At last, we formulate a global optimum cost function of the covariance matrix, by which the exact interferometric phase can be estimated when the cost function get its minimum. The method can provide accurate estimate of the terrain interferometric phase (interferogram) even if the coregistration error is more than one pixel. The real data of InSAR results show the robust and high efficiency of this method.