Three electrical engineering students from the School of Electrical Engineering and Informatics at Bandung Institute of Technology (STEI ITB) have developed SmartFishSense to determine when fish in ponds are hungry automatically.
The development of fish-related technology served as the culminating undergraduate thesis project for these three students: Agape D’sky, Oktavian Putra Masyiakh, and Alsandi Tarigan from ITB. Instead, they developed an innovation that can enhance aquaculture performance in Indonesia.
Agape, one of the members of the SmartFishSense team, explains that SmartFishSense is a fish-feeding detection system designed to monitor fish behaviour in ponds or aquaculture farms in real-time. It allows users to determine the appropriate feeding time for the fish, ensuring that the amount of feed given meets their needs and avoids wastage.
“Our device consists of several components, including an accelerometer and a webcam. These two tools serve different purposes but are necessary to detect stimuli from the fish when they require food,” Agape explained.
In this system, there is an accelerometer component. This device will be placed on the water’s surface in the pond or aquaculture farm at the location where the fish are fed. The accelerometer will serve as a sensor capable of detecting water movement caused by the behaviour of fish gathering in that area. This method is applied because fish in need of food usually gather in a specific location where the feed is provided, resulting in different water movement compared to other areas.
A camera will also be positioned above the pond. The camera’s purpose is to focus on the specific locations where the fish will receive their food. It enables precise monitoring of the feeding areas. The data collected from the accelerometer and webcam will be transmitted to a Raspberry Pi 4 Model B, employing deep learning algorithms to process the data.
Through deep learning, the Raspberry Pi can generate output data indicating whether the fish are hungry. This innovative system combines data from multiple sources, allowing for accurate and reliable detection of the fish’s feeding behaviour. By integrating advanced technologies such as computer vision and deep learning, this SmartFishSense solution offers an efficient approach to managing fish feeding in aquaculture settings.
Upon completing the deep learning process, this technology will transmit the gathered data to a server that automatically triggers the feeding device if the received data indicates that the fish are hungry.
This iterative process will continue until the fish are satiated, and the data received by the server generates an output indicating that the fish are no longer hungry. This condition will be marked by stabilisation in water movement at the feeding location, as the fish initially gathered around the feeding area disperse throughout the entire pond.
By implementing this automated feeding system based on data analysis and interpretation, the SmartFishSense technology ensures that the fish receive their food appropriately and optimally. This approach not only enhances feeding efficiency but also promotes the overall well-being and health of the fish.
Furthermore, Agape believes this system will offer a significant advantage in aquaculture management by reducing manual feeding requirements and providing a more precise and responsive feeding mechanism. “We know our technology can benefit the aquaculture industry in the future. We need more improvement to make this more accessible for the aquaculture farmers,” he reiterated.