Submarine Target Sonar Image Segmentation Based on FCM Clustering with Three-Dimensional Gray Features

Abstract

The accuracy of sonar image segmentation results directly affects the subsequent processing steps and the final understanding of sonar images, so the sonar image segmentation plays an important role in sonar application. Fuzzy c-Means (referred to as FCM) clustering can be used for image segmentation. This paper investigates the applications of FCM clustering with 10 kinds of three-dimensional gray feature vectors to sonar image segmentation. The experimental results show that sonar image segmentation using FCM clustering can obtain better segmentation results if three-dimensional feature selection is appropriate.