Improved feed utilisation in cage aquaculture by use of machine vision
Thesis (MScEng (Process Engineering))--Stellenbosch University, 2008.
With the harvesting of fish and other aquatic organisms from natural waters having reached its upper limit, aquaculture is vital in providing for the ever increasing demand for fishery products (Boyd, 1999). Not surprisingly, aquaculture has seen considerable growth over the last decade or more. With the rising importance of aquaculture, there is an increased emphasis on cost and reducing of waste for environmental reasons. Therefore, attempts to automate or increase efficiency of feeding are constantly being explored. On an aquaculture unit approximately 60% of all costs are for feed; therefore high quality feeding management is essential for all fish farmers. The rainbow trout farm at Jonkershoek Aquaculture Research farm near Stellenbosch currently have a feeding management system which makes use of traditional hand feeding. Handfeeding is not considered optimal, as the feed intake or pellet loss is not closely monitored resulting in higher operating costs. Automation of aquaculture systems will allow the industry to produce closer to markets, improve environmental control, reduce catastrophic losses, minimize environmental regulation by reducing effluents, reduce production costs and improve product quality. The history of automated control in aquaculture has been brief; most of the systems have been custom-designed, personal computer systems. A very popular approach for an automated feeding system is to monitor waste pellets beneath the feeding zone of the fish, with a feedback loop that can switch off the feeder if this waste exceeds a predetermined threshold. Other approaches use hydroacoustics to monitor waste pellets or demand feeders have also been implemented. These approaches however are not considered optimal as automatic feeders do not necessarily ensure optimal feed intake. Social dominance using demand feeders does not allow even feeding distribution among all sizes of fish. In this project it was investigated whether an automated feeding system can be developed based on fish feeding behaviour. After facing problems with poor visibility at the Jonkershoek Aquaculture farm near Stellenbosch, video data were acquired from the Two Oceans Aquarium in Cape Town. Since it was a feasibility study, the focus was rather to investigate whether a predictive model could be generated for fish feeding behaviour in a more ideal environment which can form a foundation for further research. The well-established multivariate methods of principal components