An application of the principle of minimal frustration to phase unwrapping

The problem of phase unwrapping (PU) and a new algorithm for the reconstruction of the absolute phase are presented. The PU problem is mapped to the problem of finding the global minimum of a non-equilibrium multistable system. A functional is introduced which consists of multiple constraints due to phase inconsistencies and the a priori knowledge on the regularization of the sampled wrapped phase signal. A self-annealing dynamics is used to find the state of minimal frustration of the system, where constraints on the phase signal are mostly satisfied. Experiments on a benchmark synthetic phase signal and on a real radar interferogram show the effectiveness of the proposed PU algorithm.