Phase difference based multiple acquisition phase unwrapping

Abstract

This work addresses the derivation of a phase-difference (gradient) based maximum likelihood (ML) phase unwrapping algorithm. In particular, we determine and study the structure of the ML phase unwrapping on a 2D grid in the multi-channel case and compare it with existing phase unwrapping algorithms. This allows us first to frame single-channel, phase based phase unwrapping algorithm in a general formulation. Second, among the known single-channel, phase-difference based phase unwrapping algorithms we identify those achieving a ML solution. Although achieving results similar to existing phase-based ML phase unwrapping algorithm in the multi-channel case, our approach allows to easily incorporate possible a priori knowledge about the absolute phase variation dynamics.