THIS CONTENT IS BROUGHT TO YOU BY NTNU Norwegian University of Science and Technology - read more
Researchers use AI to combat salmon lice
All models became much faster than experienced biologists at identifying the parasites that harm salmon.
Salmon lice (Lepeophtheirus salmonis) has always plagued and parasitised wild salmonids. They are one of the salmon industry’s biggest problems.
Researchers from NTNU and Wageningen University in the Netherlands have now developed a new method that could provide better control over these parasites.
They have developed large datasets that AI models can be trained on to detect salmon lice.
In a recent study, the researchers show how the method can make lice detection much more efficient.
AI was better than experienced biologists
The researchers carried out a comparison:
Biologists spent more than 30 hours over several days identifying 82 per cent of the lice larvae in one large and complex seawater sample.
The AI model developed by the researchers needed only 30 minutes to identify 97.5 per cent of the larvae in the same sample.
“A number of different measures are being used and tested to combat salmon lice. It's often the combined effect of several measures that will best improve the health of both farmed salmon and wild salmonids,” says researcher Lars Christian Gansel.
He is head of NTNU’s Department of Biological Sciences and helped develop the method.
“More information is needed about the spread of salmon louse larvae to document the effectiveness of current methods, as well as develop and customise new measures. Our model makes it possible to obtain this information,” he says.
Important to detect larvae early
Every year, between 400 and 450 million juvenile salmon and rainbow trout are released into net pens in Norway.
A single fish farm can contain millions of salmon. It can release millions of salmon lice larvae into the fjords every single day.
“If we are to succeed in eradicating salmon lice, the best approach is to prevent contact between the parasite and the fish. To develop, evaluate, and document the effectiveness of preventive methods, it's important to detect the larvae while they are still drifting around in the sea,” says the researcher.
Salmon lice are difficult to count
Compared with plankton and other particles, salmon lice can actually be considered a rare organism, according to Gansel.
This is because there are hundreds of thousands or even millions of other organisms in the sea for every salmon louse larva.
“We must therefore analyse large volumes of water in order to monitor salmon lice in the sea. If we don’t use enough water, we can easily overestimate or underestimate the number,” says Gansel.
Many methods have been used to continuously monitor and count salmon louse larvae, but the vast majority of them have been cumbersome, imprecise, time-consuming, and expensive.
Researchers created 120,000 images of lice
“Currently available camera systems for plankton analysis often lack the resolution needed to distinguish between species and developmental stages. There's still no fully documented method for continuous monitoring of salmon lice in the ocean,” says Gansel.
Artificial intelligence and machine learning have emerged as a possible solution. The challenge has been the lack of high-resolution, clear, detailed images of larvae in real seawater environments on which to train the AI models.
But the research team may have found a solution. They have built their own video microscope and taken more than 120,000 images of louse larvae and other organisms.
They have trained AI models using the synthetic data.
“The models performed just as well as the experts using microscopes. Even though some other marine species may look quite similar, the model was able to identify salmon lice in large seawater samples,” says Gansel.
Hatched salmon lice for AI training
The researchers concentrated particles the same size as salmon lice from hundreds of seawater samples.
In total, they collected and filtered several thousand cubic metres of seawater from fish farms and marine areas near Ålesund on the west coast of Norway.
Sea conditions vary depending on the season and location. When there are few salmon lice in circulation, it takes a long time to compile reliable datasets from the samples.
To create more training material, the researchers hatched salmon louse larvae themselves and released them into the water samples. They then let the water flow slowly through a glass tube while filming the particles in the water stream using a video microscope.
Researchers rescaled and rotated the lice
Using software that can track and select individual video elements, they isolated images of larvae at two different stages: newly hatched lice, or nauplii, and the slightly larger copepodites, which are ready to attach themselves to fish.
The individual frames from the videos will not show the larvae from every angle. They may be moving, or they may simply drift through certain parts of the tube.
"Because we only see a few of all the possible scenarios, we can improve the models by generating synthetic data to use alongside a large number of real-life videos,” says Gansel.
Salmon lice can also vary slightly in size.
"To account for the differences, we can scale the lice, rotate, and flip them, and include multiple lice in the same image. The same can be done with plankton and organisms that resemble salmon lice to improve the model even further,” the researcher explains.
The new models eliminate a lot of uncertainty
The models can be used to monitor salmon lice levels in areas where wild fish are expected to be present.
They can also be used to calculate the release of larvae and investigate how they spread, grow, and develop.
Monitoring can help in assessing potential measures and is important when estimating the risk of transmission between farmed and wild salmonids, the researcher explains.
“Measuring larvae directly in the sea will eliminate some of the uncertainty in the current system, where the number of larvae is estimated based on the number of lice on farmed fish," says Gansel.
This will make the salmon lice map produced by the Norwegian Institute of Marine Research much more accurate. Production can be planned more effectively,
"And we can make better decisions about where to operate fish farms and what measures to take against salmon lice,” says Gansel.
References:
Zhang et al. An image synthesis framework for enhanced salmon louse larvae (Lepeoptherius Salmonis) detection in complex seawater conditions, Computers and Electronics in Agriculture, 2025. DOI: 10.1016/j.compag.2025.110985
Zhang et al. Rapid detection of salmon louse larvae in seawater based on machine learning, Aquaculture, 2024. DOI: 10.1016/j.aquaculture.2024.741252
———
Read the Norwegian version of this article at forskning.no
More content from NTNU:
-
Can you tell the most common seagulls apart?
-
The strange hum: Are you among the few who can hear it?
-
Brains that cope well with stress resemble one another
-
Nobel prize winners with a new discovery: "It's possible to measure the inner life"
-
AI is not displacing young job seekers – yet
-
Participants became annoyed when Pepper the robot became too competitive
