A research group from the Japan Agency for Marine-Earth Science and Technology and Kagoshima University has developed an AI (artificial intelligence) system that uses images to quantify the extent of coastal pollution caused by marine debris. Until now, surveys of the pollution status have been carried out manually, so this system is likely to save labor and have an effect on the formulation of environmental policies by national and local governments.
According to the Japan Agency for Marine-Earth Science and Technology, this system works by uploading photos taken on the beach to a dedicated website, where image analysis AI detects washed-up trash and automatically quantifies the amount of trash on the beach. The dedicated website has been named "Beach LISA" and is already publicly available.
Image analysis AI has the ability to distinguish between man-made garbage such as bottles, cans, and plastic products, and natural garbage such as driftwood, seaweed, etc. from photos taken from the ground. For aerial images taken by drones, it is possible to estimate the coverage area and coverage rate of each, making it easy to quantify even without specialized knowledge.
Until now, surveys of beached litter have been done manually, requiring both money and manpower, and the methods of quantification vary from place to place, making comparisons between regions difficult. The Ministry of the Environment has added this system to its recommended tools.
The research group hopes that this will make it easier for anyone to take part in coastal conservation activities, and that it will also be effective in supporting environmental protection activities in developing countries.
