Canonical Representation of sidescan sonar images for robotics application

Abstract: Sidescan sonar (SSS) is crucial in underwater investigation as it is cheaper and provide wider detection range. Sidescan sonars are typically mounted on Autonomous Underwater Vehicles (AUVs) or vessels. However, the pixel and it’s intensity in SSS image is not corresponding to the 3D seabed perfectly, which cause some problem to utilize SSS image for navigation and mapping. This thesis is aiming to represent the SSS image to a canonical image. The approaches considered in this thesis including model fitting in range, incident angle and beam angle on the intensity in direction across AUV traveling. Furthermore, the deconvolution in direction along AUV travelling will be applied to remove the blurry due to beam width.