This is how Cermak uses artificial intelligence to create a fish health record

The iFarm farming project relies on artificial intelligence and machine learning to improve health and well-being in cages.

The technology can now identify individual fish using artificial intelligence and create a health record for each individual in the cage, according to a press release.

In September 2020, fish were released into iFarm for the first time at Cermaq. Throughout the early stages of the project, the focus was on adapting the iFarm module to fish behavior and developing the iFarm module’s camera and image quality. The goal is individual education.

The iFarm project is a collaboration between Cermaq, BioSort, ScaleAQ and Nofima. Cermaq has secured four development licenses with iFarm.

Unique individuals
Now, with the help of artificial intelligence, BioSort is able to recognize individual fish in the cage and give the fish their own health records. Basic software systems for health records and recording of individuals’ health reports are also now in operation.

– We have been working with salmon facial recognition algorithms for a long time and have built an image database of fish from different life stages, which has now been put into use online at the Hilarvika plant in Cermak in Stegen, says General Manager of BioSort, Geir Stange Hauge.

Stan Hogg explains:

– This means that every time a fish is sighted by the iFarm sensor, the fish’s identity is verified, and a health report including care indicators, lice, weight and k-factor is stored in the individual’s log for each recording. This way, we obtain a history of the individual fish that can be used to follow the individual’s growth, and any health symptoms.

Monitor the progression of the wound
The general manager of BioSort says this technology can be used to monitor wound development in fish.

For example, if a fish has a wound, we can see if it will heal or get worse, and the fish should probably be removed to avoid spreading the infection, says Stang Hogg, who adds:

– When this product is fully developed into a commercial product, monitoring of individual fish and all fish will provide a continuous summary of growth and an opportunity to document good fish welfare for the population.

Provide accurate monitoring of fish
Work is now underway to increase the database of various machine vision algorithms to increase the accuracy of individual identification and information in the fish health record. Karl-Friedrich Otten is Head of Fish Health and iFarm Project Manager at Cermak. It is believed that technology could help in keeping a close eye on fish.

In the future, machine vision for monitoring fish health will give us breeders the opportunity to follow fish more closely and, over time, we will gain a better understanding of the course of disease that we can translate into better treatment of difficult health situations, says Otim. .

Robot hoof
BioSort adopted a robotic hoof to catch individual fish and move the fish to a separate tank on the surface of the water. According to the company, this can be done without disturbing other fish and enables preventative fish health at the individual level.

– Although this is not a goal in the original project plan in the iFarm project, we still envision that the next step in development is to make the sorter, together with the sensor, operate independently and react to selected parameters, such as lice or fish with Signs of wounds or illness, concludes Ger Stang Hug.