An approach for side scan sonar acoustic images segmentation using programmable logic

Abstract

In applications related to maintenance of underwater infrastructure in harbour installations, cables and pipelines, it is necessary to separate the relevant objects of study from the rest of the acoustic image obtained by sonar. Current acoustic image processing techniques require a high computational effort and time consumption due to the manipulation of large volumes of data generated by these devices. Motivated by this problem, in this paper we present an alternative for segmenting acoustic images, proposing a framework for use with programmable logic technology. We propose to perform data segmentation into three types of regions, acoustic highlight, seafloor reverberation and acoustic shadow areas, by using programmable logic and VHDL hardware description language. Finally, results obtained over seafloor images with pipeline presence and acquired with a side scan sonar from a vessel in Salvador de Bahía, Brazil are presented.