An eigenspace-like algorithm for multibaseline InSAR phase unwrapping

In this work we present an eigenspace-like technique for the estimation of the unwrapped phase which is based on the model of the multi-baseline joint data group. The method is shown to work well even in the presence of the image coregistration misalignment and the steering vector mismatch, and has the ability to overcome the conflict associated with the computational complexity and the lack of the independent and identically distributed (i.i.d.) samples. The performance analysis of the technique is carried out based on Monte Carlo simulations and Cramer-Rao lower bounds (CRLBs) calculation. Numerical results on simulated data demonstrate the efficiency and precision of the proposed method.