Measuring feeding activity of fish in RAS using computer vision

In recirculating aquaculture systems (RAS), over-feeding events impact the welfare of Atlantic salmon and the operation of the aquaculture system. This study describes a computer vision-based method for measuring the feeding activity of an Atlantic salmon (Salmo salar) shoal. The feeding motions of Atlantic salmon were recorded with a CCD camera mounted above the rearing tank. Feeding activity analysis was based on the intensity summation of the difference frame due to the motions of the fish. An overlap coefficient was defined to calibrate the error of calculation caused by the overlaps among fish bodies in images. An approach to filter out the influence of water surface reflections on preceding results was proposed. Based on these data, a computer vision-based feeding activity index (CVFAI) was determined for measuring the feeding activity of the fish in arbitrary given duration. To assess the reliability of CVFAI, a manual observation feeding activity index (MOFAI) was determined by scoring each kind of feeding behavior in the same recordings. The CVFAI and MOFAI presented a linear relationship at a correlation coefficient of 0.9195. CVFAI is therefore a potential indicator for measuring the feeding activity of an Atlantic salmon shoal, in a low-cost and rapid way. (C) 2014 Elsevier B.V. All rights reserved.