Parametrical and Textural Analysis of Sidescan Sonar Images of the Seafloor

This paper presents a comparative analysis of a new method of sidescan sonar imagery segmentation helpful in the estimation and spatial distribution of macrophytobentos in Arctic conditions. In our research, we designed three methods, two based on a parametric analysis of sidescan sonar echo signals and the third on a textural analysis relying on mathematical morphology. Acoustic observations were verified by video recordings and biological samplings. A single beam echosounder was also used. The Spitsbergen fjords represent a periglacial environment with great diversity of morphodynamic processes and sensitivity for global warming changes, so it is one of the most promising areas to study the influence of climate change on an ecosystem. Sidescan sonar is a very effective and economic tool for mapping marine vegetation on the seafloor, but interpretation of data still causes many problems especially in the specific conditions of the Arctic fjords (underwater rocks, postglacial sediments, steep slopes). Proposed segmentation and classification algorithms enable estimation of the bottom area covered by algae and, together with results of biological sampling and single beam echosounder measurements, estimation of algae biovolume and biomass.