Intelligent Classification of Sonar Images

In many research areas, intelligent recognition and classification systems gained an important role. The reliability and the success of these systems are depend on the effectiveness of applied data preprocessing techniques and neural networks which can be used for efficient modeling of human’s visual system during the recognition or classification of patterns. Neural networks have an important part in the modeling of human experience and decision – making process into computers. In this paper, Sonar Image Classification System which was developed to simulate human experience in the recognition of underwater shapes by using Pattern Averaging and Back Propagation Learning Algorithm, will be presented. Experimental results suggest that automatic intelligent classification of these shapes may provide more effective researches in oceanic engineering. Key-Words: Pattern Averaging, Neural Networks, Sonar Images.