Phase ambiguities in Synthetic Aperture Sonar Interferometry

Synthetic aperture sonar (SAS) interferometry is a technique used to make very high-resolution sonar images and bathymetric maps of the seafloor, while simultaneously achieving a relatively high mapping rate. Traditional SAS and synthetic aperture radar (SAR) interferometry techniques rely on coregistered images and advanced phase unwrapping techniques to resolve phase ambiguities. For wideband systems, the absolute (or ambiguity-free) phase difference between the interferometric images can be estimated directly. However, since the phase difference is a stochastic variable with a multi-peak probability density function (PDF), even the absolute phase difference estimators will experience wrap errors. In this paper, we study the probability of wrap errors as a function of signal-to-noise ratio (SNR), bandwidth and the number of independent samples used in the phase difference estimator. We suggest two methods to reduce the amount of wrap errors in the interferogram: A local median unwrapping filter and a two-stage unwrapping method. Both of these methods use a larger, but local region of data, to unwrap the phase. We verify the theoretical performance of the interferometric processing on simulated data and demonstrate the suggested unwrappers on data collected using a HUGIN autonomous underwater vehicle (AUV) equipped with a HISAS 1030 wideband interferometric SAS.