Bayesian approach to phase-difference-based phase unwrapping

Abstract

We address the study of phase-difference-based phase unwrapping in a statistical framework. We derive the joint statistics of the measured phased differences on a 2D grid; for two cases, we measure a single or several (scaled) wrapped value(s) of the absolute phase. We then determine and study the structure of the unwrapped phase maximum likelihood estimator and compare it to existing 2D phase-difference-based, single-acquisition, phase unwrapping algorithms. Furthermore, we investigate the possibility of introducing a-priori information.