Image
AI algorithm identifies fish, highlights them in green, and estimates their weight over time.
Artificial intelligence is used to identify fish in order to estimate their weight. The image shows an example of how the algorithm highlights fish in green to track individuals over time.
Photo: IVL, Svenska Miljöinstitutet
Breadcrumb

How AI is helping to improve fish farming

Feeding fish might sound simple but in aquaculture every pellet counts. Researchers at the Swedish Mariculture Research Center, SWEMARC, and project partners are now improving fish farming using Artificial Intelligence (AI) to make sure fish get exactly what they need, when they need it. We asked Jonathan Roques, docent at the Department of Biological and Environmental Sciences at the University of Gothenburg, to explain how AI can be used in modern aquaculture.

What are the main advantages of using AI in fish farming?

"Some procedures are labour intensive, like feeding or estimating the biomass and the welfare of your fish. When you have thousands of fish in a tank, you cannot check them one by one, you rely on sub-sampling which may not be fully representative." 

"If you install an automated system where you can monitor many individuals at the same time, it is beneficial for the farmer, and the fish. You have a live overview of what is happening in your tanks, and you can better plan your operation and act quickly if you notice something strange."

How does AI assisted feeding work in practice?

"In our project, we use a combination of camera and hydrophones to look at and listen to the fish to monitor their activity. We feed this data to the AI, so it can predict when the fish are hungry and when they are full to know when it is time to start and stop the feeding."


What are the main benefits for fish farmers?

"Feed is the main cost for the farmers, it is important to provide your animals with the right type of feed, which is another area we are working in within SWEMARC, and it is also important that the fish are fed the right way. We want to avoid suboptimal feeding, both underfeeding and overfeeding. The farmer wants the fish to grow well (and fast), so the fish should not be underfed. However, the feed should be converted into muscle, not visceral fat, so they should not be overfed." 

"In addition, uneaten pellets are a waste of money, and they can also create issues with water quality, impairing your system and potentially fish health and welfare. Currently, many farms are equipped with monitoring systems for water quality parameters for instance, and there is research done on how to improve feeding and welfare check."


How do you see the role of AI in the future of aquaculture?

"AI will be a great tool to help farmers with their daily activities. It will facilitate their operation and allow them to monitor more parameters in real-time in an objective way, help plan their activities and give them early warnings. It will save them time and money, but it should not be seen as a replacement for their expertise. Ultimately, they will still have the last word and make the decisions. Farmers should still have a good knowledge of their system and their fish."

Facts

Project name: Sustainable Industrial Production of Fish with AI and Advanced Data Analytics - Implementation (HIFAI)

Project period: February 2023 to December 2025

Project partners: IVL Swedish Environmental Research Institute, University of Gothenburg, Innovatum Science Park, Gårdsfisk, Smögenlax.

Read moreHIFAI - IVL.se