Simulation based algorithm for tracking fish population in unconstrained underwater

Abstract

Tracking of object in underwater is a emerging idea. There are several type of approaches used for this but they have high computational complexity. But scientist needs to study fish populations underwater in different environmental conditions. In this paper we proposed a simple algorithm for tracking and detecting multi moving fishes. A huge number of fishes can be tracked by comparing its mean in the successive frame of the video. The obtained output states a measurement of the difference of two mean values. The object window contains the matrix of the intensity value. Then this value is transformed into a set of feature vector. Finally the above two set of features is compared in the successive frame for a good match to its nearest locality. The MatLab 8.0 simulation result shows that the proposed method is capable of accurately detecting with high detection rate as compared to existing approaches with noisy and dense environment.