Artificial Intelligence Joins Fight Against Marine Litter
MARLIT, an open-access web app based on a deep-learning artificial intelligence algorithm, will promote the detection and measurement of floating plastics in the sea with a reliability of over 80 percent, according to a study published in the journal Environmental Pollution conducted by experts at the Biodiversity Research Institute of the University of Barcelona.
Floating macro-litter is a threat to the conservation of marine ecosystems worldwide. The largest density of floating litter is in the great ocean gyres (systems of circular currents that spin and catch litter), but polluting waste is also abundant in coastal waters and semi-closed seas such as the Mediterranean. The analysis, using artificial intelligence techniques and more than 3,800 aerial images of the Mediterranean coast along Catalonia, will permit researchers to detect the presence, density and distribution of plastic pollutants in seas and oceans worldwide. Historically, direct observation by boats and planes formed the basis for assessing the impact of floating marine macro-litter, but the massive area of the oceans and the amount of data required have made it hard in the past for researchers to progress with new monitoring studies.