A Sensor-Actuator Network Towards the Optimization of Feeding in Marine Aquaculture

As the aquaculture industry continues to expand, optimization of feeding continues to be a crucial issue in mitigating the environmental and economic impacts of poor feeding management, improving the growth and health of the fish, and keeping the industry sustainable. Recognition of fish-feeding behavior and control systems are important tools for estimating the optimal amount of food to be dispensed. In this paper, we proposed a sensor-actuator network made for fish-feeding behavior recognition using fish-induced flow speed measurements and underwater fish imaging. This system also consists of a feeding control module, a motor control panel board, and a feeding machine to control the feeding. We deployed this system at a fish farm for one month. After collecting data, we trained a model to detect the fish from the video recorded near the surface and compared the results from the flow speed measurements. Similarities between the patterns of the fish count and the flow speed showed some promising results that can be used to optimize fish feeding.