Production of fish in land-based aquaculture contributes significantly to the production of animal protein in the human food-supply chain in the Nordic countries. Recirculation aquaculture systems (RAS) is an environmentally sustainable solution that is becoming increasingly relevant for land-based aquaculture. However, production and health management are considerable challenges which limit production performance.
In most RAS systems, production and health management relies on visual observations of the behaviour of the fish (feeding activity and detection of abnormal behaviour) in absence of more automated systems. Today, farmers need to invest a great deal of time to observe the fish, and a large amount of training is required to become familiar with typical- and atypical behaviour of the fish.
An extensive amount of data on water quality, feed use and health parameters are gathered through monitoring of the fish and their environment via sensors. Thus, each farmer has access to a vast amount of data originating from their own farm. In most cases, these data are stored in separate data systems, not utilizing the huge potential for more precise monitoring and reporting of the realtime health, welfare and growth of the fish that could be achieved if data were combined.
In this project we will use Artificial Intelligence (AI) and statistical models to aid the transition from a human experience-based management of RAS production to a knowledge-based automatic one. Feed management, feeding and feed waste is a major challenge to production. In marine aquaculture, video systems are widely used to observe the fish. This has not yet been implemented in RAS systems. Within this project we will apply deep learning tools on recordings of thousands of video sequences of fish, teaching computer systems how the fish react and use this in feed management and health monitoring. We aim at developing a closed loop control feeding system to optimize growth and limit feed waste and at the same time explore the possibilities of an early detection system for health monitoring.
The results from this project will be directly applicable in RAS farms everywhere and will improve fish health and welfare as well as productivity, thus ensuring an increasing sustainability of aquaculture.