Automated Optimal Processing of Phase Differencing Side-scan Sonar Data using the Most Probable Angle Algorithm

Phase-differencing side-scan sonar systems produce co-located bathymetry in addition to each side-scan amplitude measurement. Bathymetric soundings are calculated from the range to each measurement (derived from the two-way travel time) and the receive angle of the incoming signal. Because phasedifferencing systems produce a seafloor sounding with each individual measurement, they are often characterized as noisy when compared to multi-beam sonar systems, whose seafloor estimates, whether by amplitude-weighted mean or sub-aperture phase difference detection, are the product of averaging several measurements. In addition, every effort is made to increase the resolution of side-scan data by increasing the bandwidth and sampling rate of the transmitted signal, often producing more than 10,000 data points per ping. This volume of outlier-prone, relatively noisy data is difficult for operators to interpret and software to process. A series of methods has been developed for the automated processing of phase-differencing side-scan sonar data producing seafloor estimates and related uncertainties optimized for the survey application. The “Most-Probable Angle Algorithm” (MPAA) has been developed for the filtering of outliers in range-angle measurements. With outliers removed, the uncertainty of the filtered measurements are estimated. Angle estimates are then calculated as an uncertainty-weighted mean where the number of measurements contributing to each estimate is determined from that required to achieve a desired depth uncertainty. The resulting swath of depth measurements contains irregularly spaced soundings, typically obtaining full spatial resolution of the side-scan data from 20-50 degrees from nadir, and combining several measurements to reduce the uncertainty elsewhere. In this way, given a survey requirement, an optimal amount of information can be extracted from the sonar data in varying conditions