Reveals fish larvae’s health in no time

April 29, 2019 Facebook Twitter LinkedIn Google+ Aquaculture News,News-Europe

Researchers can now make lightning-fast analyses of growth and health for small-sized marine species, thanks to artificial intelligence. The system has reduced the analysis time from 10-15 minutes to 2-3 seconds.

 – This is very useful in the lab and will save us for many research hours. More importantly, it is out in the field and at fish farms where it can provide a real-time update of the fish fry’s growth, says the SINTEF researchers behind the new technology.

They have connected available software with a proprietary underwater microscope camera called SilCam. Now they can take photos and analyse them quickly.

– With the system we will also be able to find patterns from data that is difficult for people to discover, says researcher Emlyn Davies at SINTEF Ocean.

Together with SINTEF researcher Bjarne Kvæstad, he has trained the system to look at how fish larvae react to environmental impact.

– We have set targets for fresh fish larvae and examples from experiments carried out at the lab where they were exposed to environmental toxins, says Kvæstad.

By targeting the fish larvae, researchers can see if they are normal or deformed.

Saves a lot of time

Previously, this process had to be done manually, and it usually took 10-15 minutes per larva, but with the new system, the analysis can be done in only 2-3 seconds.

– A project manager saved his project for around 150 research hours when he was recently analysing 750 pictures of fish larvae. The PC did the analysis job in half an hour. It can also be mentioned that this is a question of boring boring work, which also requires expertise, says Kvæstad.

The system makes it much easier to go through a lot of data in a short time, and can lead to research that would otherwise be too resource-intensive.

More applications

“So far, we have tested the new system on fish larvae, but we imagine that it can also be used on other different marine species, such as jumping crabs and brushes,” says Davies.

The system has so far been used in labs, but the researchers believe it is well suited in fields where it can have great significance for knowledge and management of the sea, as well as the development of land-based farming.

“This can be used for real-time measurement of breeding breeding in breeding plants, and when we get it mounted on autonomous vessels, we can map information in the sea about everything from fish eggs and phytoplankton to microplastics,” says research director at SINTEF Ocean Atle Kleven.

The researchers also have a goal of automating the system so that it can be used for environmental monitoring, for example in the case of oil spills.


Photo caption: The pictures show a normal and a deformed fish larva. The deformed caterpillar differs from normal because it is shorter. It also has cranial, jaw and spine deformities, and both plum sack and heart edema. Photo: Bjørn Henrik Hansen, SINTEF Ocean