Ensuring efficient fish passage: mapping fish using AI
To ensure that the fish passage works efficiently, Vattenfall initiated a project to test how artificial intelligence (AI) can be used to monitor the effectiveness of fishways.
Short facts
- Stornorrfors hydropower plant, Northern Sweden
- Salmon and trout
- Efficient fish passage and monitor fish population health
- Continuous
An AI program was developed and trained on a substantial image dataset to directly analyse images of migratory fish captured underwater.
The AI-based fish recognition system is deployed at the fish ladder in the Stornorrfors hydropower plant in northern Sweden, where it counts and registers attributes of individual salmon and trout. The algorithm has been progressively refined each season and now provides real-time fish monitoring data through an online dashboard.
The latest enhancement allows for measuring the size of fish, facilitating pattern visualization on a heatmap and assessing the health of the fish population. The collected data also offer insights into the behaviour of the fish, such as swimming patterns, group dynamics, and gender identification.
By analysing this data, assertions can be made about the state of the fish population and strategies for its improvement can be devised.
More about biodiversity
Biodiversity and nature protection are a priority at Vattenfall. It is one of the focus areas in our environmental policy and therefore also a central part in our environmental work.
See also
Ensuring safe fish passage: guiding fish using pumps
At large hydropower plants, guiding fish is complex and conventional intake racks are often too costly or impractical.
Biotope restoration and species protection
Biotope restoration and species protection play a vital role in adapting hydropower to the environment.
Enhancing marine biodiversity: nature inclusive design at Hollandse Kust Zuid
Nature-inclusive offshore wind design adds features to support native species.