How Fish Farms Can Use Facial Recognition to Survey Sick Salmon
A Norwegian aquaculture company plans to combat sea lice and other problems by monitoring individual salmon in a high-tech fish farms
Facial recognition tools are used in criminal investigations and when you tag your friends in photos on Facebook, but now, its usage is getting a little fishy—literally. A fish farm company hopes to use face identification technology in high-tech salmon farms to check the animals for a parasite called sea lice and other health problems.
The Norwegian fish-farming firm Cermaq Group AS is planning on implementing the facial recognition software as part of a high-tech fish farming suite they are calling iFarm, reports Agnieszka de Sousa at Bloomberg Businessweek. The tech, which should be ready for commercial use in five to six years, is likely going to be in high demand.
Currently, most fish farms assess the health of their salmon as a group, not individually. If a few fish are found to have a disease or parasite, the whole farm is treated. But the iFarm system aims to assess each individual fish, allowing fish growers to determine how fast each fish grows and check for the presence of disease or parasites.
To do this, the system will use a machine called the BioSort vision recognition system. De Sousa at Bloomberg reports that the system relies on the biology of the fish. In an iFarm, about 200,000 salmon would live in 525-foot circular net. About every four days, salmon need to come to the surface to take a gulp of air to regulate their swim bladder. When they do, a pyramid shaped device guides them into the camera that recognizes their face based on the pattern of dots on their snout and gills and also scans their entire body. If the fish shows signs of a problem, it is then guided into a holding pen for individual treatment.
The iFarm tool is especially helpful in dealing with sea lice, a parasite which has become huge problem in the salmon industry, infecting fish farms in Norway, the U.S., Canada, Chile and Scotland. The lice, which are a type of crustacean, cost the salmon farming industry $1 billion per year.
Farmers are trying all sorts of techniques to rid their fish of the lice, including dousing them with pesticides, breeding genetically resistant fish and even using lice-zapping lasers, reports Patrick Whittle at The Independent. Others are experimenting with raising certain fish species that will slurp the lice off salmon. Some are using a technique that raises the temperature of the water briefly to cause the lice to detach. But the iFarm system would allow farmers to be more targeted in their approach.
“We know that sea lice are very unevenly distributed amongst the fish, and this system enables us to avoid mass lice treatments,” Cermaq Norway Managing Director Knut Ellekjær tells Nicki Holmyard at Seafood Source. “Similarly, we can sort salmon on the basis of weight and remove only those fish ready for harvest, without stressing the others.”
It’s estimated that the system could cut fish mortality from sea lice by 50 to 75 percent. Even when lice don’t kill the fish, the salmon are still rendered unsellable because of the lesions and sores the lice cause.
“Only the fish that actually need it will be sorted out for treatment, which means typically 5 to 20 percent,” Geir Stang Hauge, CEO of BioSort, the tech company collaborating with Cermaq tells De Sousa. “This avoids stressful treatment for all the healthy fish.”
It’s hoped that such a system would help bring down the price of farmed salmon, which has become more expensive as the lice spreads.
Salmon aren’t the only species getting special treatment using facial recognition. The agribusiness giant Cargill announced earlier this year that it is investing in the Irish software company Cainthus. Using image monitoring, Cainthus software is able to track individual dairy cows and monitor their food and water intake as well as their behavior patterns to determine any signs of disease or illness.
Facial recognition is also being used in animal conservation. Last year, researchers at Michigan State University created a system that could recognize the faces of individual lemurs. It’s hoped that, with some refinements, such systems could help count, track and even keep tabs on the health of endangered species in the wild.