Multi-channel model for sonar image segmentation

This work deals with unsupervised segmentation of images supplied by high resolution Sonar. Image is segmented into three kinds of regions: echo, shadow, and sea-bottom reverberation. Sonar image is passed through a bank of Gabor filters and the filtered images that possess a significant component of the original image are selected. The selected filtered images are then subjected to a non-linear transformation. An energy measure is defined on the transformed images in order to compute texture features. The texture energy features are used as input to k-means clustering algorithm. Results of the proposed method are presented for different sonar images to demonstrate the robustness of this method.