Japanese researchers develop AI system to detect marine plastic waste

Japanese researchers have developed an artificial intelligence system to better detect marine plastic waste on the seabed, according to local media Thursday.

Researchers at the Japan Agency for Marine-Earth Science and Technology developed a system called DeepLitterAI that can process data at almost twice the speed of humans, Tokyo-based Kyodo news agency reported, citing the journal Environmental Pollution.

The system is expected to be used for real-time monitoring of marine plastic waste.

Researchers created a dataset of about 12,000 images based on seabed footage captured by the agency since 1983. The images featured small litter objects as well as non-litter objects, such as rocks and marine creatures, that can easily be mistaken for waste.

They trained the system using various patterns, including blurring and image inversion, to reduce false detections.

While much of the plastic waste that enters the sea sinks to the seabed, accurately detecting it remains difficult.

"We can quickly identify areas where large amounts of waste accumulate and use this information to implement countermeasures," said Ryota Nakajima, a biological oceanographer on the research team.

During tests using actual footage, the system identified the type and quantity of litter even when objects appeared as small as 5% to 10% of the image width.

It successfully detected 80% of major litter items, including plastic bottles and polythene bags, while the margin of error compared with visual inspections by experts was about 10%.