By Mette Lynge Hansen, Independent Research Fund Denmark
Ocean warming, higher water levels, waste pollution and declining biodiversity. The world’s oceans are increasingly suffering from climate change and environmental problems. But to understand the relationships, we need to know more about what's going on beneath the ocean surface.
And that is a challenge with the methods we have today. The earth is covered by more than 70 percent water, so it is neither possible nor profitable to manually measure factors and analyse conditions over a long period of time.
Using computer vision, the research project Marine Analytics will develop artificial intelligence algorithms for a new system that automatically analyses the underwater conditions in waters with varying visibility as in Europe.
The project is supported by the Independent Research Fund Denmark and headed by Thomas B. Moeslund, Professor at Aalborg University. The study is being conducted in conjunction with the American company Kitware and the Danish company Ambolt.
- Automation continuously generates data on what is happening below the surface of the ocean. The more data the better. This will save time and money, but it will also provide a larger and better amount of data than what we’ve had thus far, says Thomas B. Moeslund.
OCEAN LIFE UNDER THE LIMFJORD BRIDGE
Three underwater cameras have already been installed at a depth of nine meters on the Limfjord bridge between Aalborg and Nørresundby. Researchers at Aalborg University can view the live feed and will now automatically monitor what is going on in the water.
- We need to investigate how visibility can be automatically determined from the images and what impact this has on the subsequent analysis. With artificial intelligence, we will be able to find and classify by species the different types of biological organisms that are now moving past the cameras, says Thomas B. Moeslund.
Specifically, the researchers will develop three algorithms. The first one looks at the quality of the images. If the quality is high enough, then another algorithm takes over to see if anything abnormal is going on in the images – anything other than the seabed with a little seaweed – such as a crab, fish, plastic or other foreign matter. If something abnormal is happening in the image, then the third algorithm takes over to figure out what the abnormal occurrence is.
Initially, the algorithms will be used by marine biologists. But the overall goal of the project is that the artificial intelligence can help gather more and better maritime data on the world's underwater conditions so that environmental problems can be understood and remedied in the future.
The project will run over the next three years.