Sand ripple characterization using an extended synthetic aperture sonar model and MCMC sampling methods

Side-look synthetic aperture sonar (SAS) can produce very high quality images of the sea-floor. The aim of this work is to develop a hierarchical Bayesian framework for estimating sand ripple characteristics from SAS imagery that can make use of multiple passes over an area at a variety of ranges. Using a hierarchical Bayesian framework and given a known sensing geometry, a method for estimating bathymetry and sand ripple frequency values from single SAS images as well as sets of SAS imagery over an area is presented. This is accomplished through the development of an extended model for sand ripple characterization and a Metropolis-Hastings sampler to estimate bathymetry and sand ripple frequency characteristics for multi-aspect high-frequency side-look sonar data. This hierarchical Bayesian framework allows prior information obtained from previous passes over an area to aid in refining bathymetry and frequency estimates through the use of prior distributions on these values. Results are presented on synthetic and real SAS imagery that indicate the ability of the proposed method to estimate desired sand ripple characteristics.