Data Driven Methods for Deriving Bathymetric Maps from Side-Scan Sonars
Accurate higher resolution bathymetric mapping currently requires multibeam echo sounders, MBES, and excellent navigation. Therefore it is carried out by ships or large AUVs. A hand launched AUVs can not carry the type of MBES needed but they can carry rather good quality sidescan sonars. The challenge is to then extract the missing spatial information from the sidescan signal which gives the intensity, in a ‘plane’ and at a range from the sonar. Furthermore, thenavigation accuracy of a small AUV will be far worse than from a surface vessel. We propose to use data driven methods and simultaneous localization and mapping, SLAM, to address these issues. We will show the progress we have made towards the goal of accuate mapping by teams of smaller AUVs.