Sonar images classification of seabed physiognomy based on the information fusion methods

Abstract

As the exploration of the seabed is extended ever further, automated recognition and classification of sonar images become increasingly important. However, most of the related methods ignore the directional information and scale invariance or only pay attention to one of them. To deal with this problem, we combine 2D Gabor filters and fuzzy fractal dimension based on the information fusion method in this paper. The 2D Gabor filters are designed with constrained parameters to reduce the complexity and to improve the calculation efficiency. Meanwhile, at each orientation, the optimal parameters of the Gabor filters will be selected with the help of bandwidth parameters based on the Fisher criterion. Hybrid-fusion method and multilayer perceptron classifier are applied in the information fusion method. The approach proposed in this paper can overcome some disadvantages of the traditional approaches of extracting texture features, and improve recognition rate effectively. ©2008 IEEE.