New Interferometric Phase Unwrapping Method Based on Energy Minimization From Contextual Modeling

Abstract

Two-dimensional phase unwrapping is a critical processing procedure of synthetic aperture radar interferometry. This operation becomes even more difficult in the presence of noise. Both local and global unwrapping algorithms are used in obtaining the absolute phase from the wrapped ones. However, there are still gaps in the use of these methods. Local methods still let noise propagate. While the global methods are unable to properly formulate the objective function. The aim of this article is to present a new algorithm for interferometric phase unwrapping. An alternative method to consider both local and global constraints of the interferogram is proposed. Local constraints are modeled by a novel quality measure. A weight is then assigned to each pixel of the interferogram based on two criteria—the phase jump ratio and the phase gradient. While the global constraints are modeled by a cost function that calculates the sum of the pixels weights in the interferogram. This cost function is minimized by a genetic algorithm. Validation tests with real and simulated interferograms showed that the proposed algorithm outperforms standard algorithms of phase unwrapping such as branch cuts, minimum cost flow networks, or minimum $L^{p}$ -norm algorithm.