Detection and counting of uneaten food pellets in a sea cage using image analysis

Abstract The purpose of this study was to detect and count feed pellets in a sea cage using underwater video cameras. Using a light-compensating camera pointing straight down in the water column, extruded pellets appear white. This effect made it possible to detect and count feed pellets during a feeding event. The manual counting of food pellets from video replay is laborious so algorithms were developed for detection and counting of food pellets from recorded video image sequences. The algorithms were implemented on a personal computer based image processing system. Experiments were performed to test the algorithms with pellet densities used in actual feeding situations. The average count error was approximately ± 10%. By increasing the video sampling rate and screening off the cameras from the fish, this error could be significantly reduced.