JAMSTEC develops AI to detect ocean floor plastic waste

Researchers at the Japan Agency for Marine-Earth Science and Technology (JAMSTEC) have developed DeepLitterAI, an artificial intelligence system designed to automatically detect, classify, and count macro-sized waste on the ocean floor. The system addresses the difficulty of manually reviewing hours of underwater video footage. In tests, DeepLitterAI analyzed waste approximately 2.1 times faster than humans, reaching up to 3.1 times faster in certain conditions. The AI was trained using the ‘J-Litter’ dataset, consisting of 12,029 images, to improve the detection of small or distant objects that traditional models often miss.

Separately, experts from Hasanuddin University have warned that approximately 90 percent of land-based waste has the potential to contaminate oceans and become a source of microplastics. This waste typically enters marine environments through canals, drainage, and rivers. Large plastic waste (macroplastics) that is not properly managed can fragment into microplastics due to weathering, water movement, and temperature exposure. Experts emphasize the importance of waste sorting at the source to facilitate recycling and prevent environmental degradation.