Author(s): S. Abdelaziz; A. Al-Nuaimi; P. Chugh; M. D. Bui; P. Rutschmann
Linked Author(s): Peter Rutschmann
Keywords: Fish path tracking; Image processing technique
Abstract: Understanding the fish behavior as a function of hydraulic conditions is the key for a good design of the fishway and other fish barriers. For that we need a method to extract the fish movement over time and relate it to changes in flow velocity, depth and other hydraulic conditions. In this study, video (image) processing techniques are applied to track the fish movement in a controlled test environment. Here we specifically use the technique motion detection through background estimation and subtraction. We further improve the results. Furthermore we present a color-histogram based segmentation approach to track the identified fish over time and hence achieve robust fish tracking. Finally we apply signal processing techniques for post processing of the produced movement curves to filter out outliers. The algorithms are applied on two test datasets, one of them captured in the Laboratory of Hydraulics in Obernach of the Technische Universit? t Munchen (TUM). Initial results are very promising and a number of improvements are furthermore suggested.
Year: 2012