Abstract: It is difficult to extract the features of underwater acoustic images because of the noise interference, multi-pass, reverberation. A method of target from sidescan sonar image auto detection is summarized by considerable experiments and analysis, including image enhancement, binary, binary opening and binary closing, feature extraction, feature computing and target recognition.

Key words: binary morphology, binary labeling, side scan sonar, target detection, median filter, Otsu binary segmentation