Sonar Image Segmentation based on an Improved Level Set Method

Abstract Aiming at the problem that the existing image segmentation methods cannot be accurately applied in sonar image segmentation, an improved level set sonar image segmentation method was proposed in this paper. Firstly, this paper analyzed the different characteristics of sonar image from the optical image, and the biggest drawback among them is the existence of shadow interference, namely for a sonar image with shadow part, the traditional level set segmentation algorithm will often make the shadow as the segmentation target to be exported out because the feature of sonar target object is not significant enough; Secondly, to overcome the shadow negative effects in sonar image segmentation and achieve selective segmentation, this paper did sonar image preprocessing by morphological top-hat and bottom-hat transformation, then carried on level set method without re-initialization and constructed an improved level set sonar image segmentation system; finally, compared the improved level set method with the traditional level set method in the simulation experiment, and the results showed that the improved level set segmentation method is more adapted to sonar image with uneven background.