Fuzzy logic based obstacle detection in images captured by imaging sonar

In this paper, an obstacle and target detection system for safe operation of unmanned underwater vehicles is proposed. An imaging sonar is used to scan the underwater environment, and a novel pre-processing method is suggested to eliminate the noise introduced by surface reflections and other reasons in the acoustic images taken by the imaging sonar. Afterwards, fuzzy logic based target/obstacle detection is utilized using the noise reduced sonar images, and experimental results are presented for different scenarios.