A Closed-Form Cluster-Analysis-Based Multibaseline Phase-Unwrapping Algorithm

Abstract

Multibaseline phase unwrapping (PU) is a critical step for the synthetic aperture radar interferometry (InSAR). The cluster analysis (CA) algorithm is a fast multibaseline PU method proposed by H. Yu. In this paper, a closed-form cluster analysis (CFCA) algorithm is presented to eliminate the ghost clusters in the CA algorithm. In this method, information of the ambiguity vector is used to identify the ghost clusters and real clusters, and the ambiguity vector of each cluster is solved by a closed-form solution, then a special neighboring pixel mean filtering measure is adopted to get a better result. Simulation results are shown to verify the effectiveness and noise robustness of the CFCA multibaseline PU method.