Smart tools for fish farming tested in new study
Imagine being able to know when the fish in a land-based farm have eaten their fill and when they have reached slaughter size. With acoustic aids, technology has been developed that works on a laboratory scale, but transferring it to a production environment is another matter. Now the first 'Blue Food' application has been given the green light and the project team led by Torbjörn Johansson at IVL has nine months to show that sound is the way forward.
In today's land-based fish farms with water recirculation, the amount of food the fish will eat is estimated and the feed is distributed over a given period of time. However, the appetite of the fish depends on factors such as water temperature and oxygen levels and can therefore vary. What is not eaten ends up in the water, breaking down and leading to a deterioration in water quality. Feed also accounts for a large part of both the cost and climate impact of fish farms, so there are strong arguments for getting the feeding right.
We will study two methods, says Torbjörn Johansson. The first is to listen to the eating sound - pellets being crushed - or the sound of the eating phrenzy that occurs when the fish are fed. When the fish start to get full, the sound level decreases. We will start with tests in the University of Gothenburg's test facility and then move on to one of Gårdsfisk's facilities. The second method is to use video and record the eating behavior of the fish, which is interpreted with the help of artificial intelligence.
For both methods, there are plenty of questions to answer such as whether the right sound can be distinguished, whether an existing artificial intelligence can be reused, whether extra light is needed and, for example, whether reflections on the water surface can be a problem.
Submarine sounds measure fish size
The second part of the project involves knowing when the fish start to reach the right size and are ready for slaughter. For this, the researchers will use high-frequency pinging sound signals - similar to those known in popular culture from submarine movies. The acoustic sound propagation is affected by the biomass of the culture system and can indicate both the total amount of fish and the size of the fish.
A third part is about optimizing feeding based on the oxygen content and temperature of the water. In all cases, the aim is to investigate and demonstrate whether the technology works in a production environment, known as 'proof of concept'.
Finally, the results will form the basis for a new application for a future major research project to develop a concept that can be used in industry. This also includes identifying which companies may be interested in participating in a collaboration. Torbjörn Johansson believes that it should be possible to have a finished product within five years. The economic and environmental incentives are strong, and previous experiments have shown that the amount of feed can be reduced by 30 percent without impairing the growth of the fish.
Funds from the Strategic Innovation Program for the process industry
The new study has a long name: "Sustainable industrial production of fish using AI and advanced data analysis" and the funding comes from a perhaps unexpected source. "Process industrial IT and automation - Piia" is a so-called strategic innovation program to promote innovation in the process industry, with Vinnova, Formas and the Swedish Energy Agency as the sender. When the project group heard that the area of fish farming was discussed during a Piia workshop, they decided to apply for funding from there.
Project leader Torbjörn Johansson, a researcher at IVL, is assisted by Kristina Snuttan Sundell, professor of zoophysiology at the University of Gothenburg, and Lillemor Lindberg, innovation manager at Innovatum Science Park, who, like Torbjörn, are active in Blue Food. Blå mat also contributes to the financing of the study and one of the companies participating in the study, Smögenlax, is also a partner in Blå mat.
For Torbjörn Johansson, an expert in underwater sound and signal analysis, collaborating with researchers who are experts in living organisms has been a new experience. But he notes that it is in the meeting between disciplines and between academia and industry that there are great opportunities to take steps forward. In other words, blue food has been a good starting point for the new study.
The picture shows two autonomous hydrophones, which will be used in the study. Photo: Torbjörn Johansson, IVL Swedish Environmental Research Institute
Facts and figures
Project title
"Sustainable industrial production of fish using AI and advanced data analysis"
Project manager
Other project participants from Blue Food
Participating organizations from Blue Food
The company Gårdsfisk and other researchers from IVL and the University of Gothenburg are also participating in the project.
Project period
October 2021 to July 2022
Financing
SEK 599,150 from Vinnova, Formas and the Swedish Energy Agency's call for proposals "Process industrial IT and automation"
SEK 270,000 from Blå mat
SEK 575,600 own funding from the participating organizations
Total budget: SEK 1,444,750.