A fast sonar-based benthic object recognition model via extreme learning machine

The fast sonar-based object recognition turns out to be one of the most challenging topics in the underwater signal analysis. In this paper, we try to develop a fast benthic object recognition model via the extreme learning machine (ELM) on the basis of the structured geometrical feature extraction. Geometrical features such as major and minor axis, eccentricity, circularity and so on are employed to construct learning samples of ELM. The classifier based on ELM is used to recognize the target objects in sonar images. It has been shown in the simulation experiments that the proposed model could keep a quite good recognition performance with a much fast speed.