Sar Phase Unwrapping Using Adaptive Recursive Smoothing

Regarding the interferometric evaluation of Synthetic Aperture Radar (SAR) data, phase unwrapping is one of the most difficult processing steps. In the past several methods have been developed, such as the least squares or the branch cut methods. In this paper the influences of terrain shape and phase noise on phase unwrapping are discussed. A new approach to phase unwrapping is taken which consists of preprocessing the interferogram before the phase information is unwrapped: the interferogram is recursively smoothed, and the differences of consecutive smoothed versions of the interferogram are computed. The phase data of the difference interferograms and the final smoothed interferogram are unwrapped, and the unwrapped phases are added. The result of the process is the unwrapped interferometric phase. It is shown that the approach largely avoids the problem of ambiguous integration paths.