AI Technology Confirms Maldives Successfully Eliminated Burning Smoke and Plastic Pollution
The Republic of Maldives faced significant environmental challenges in early 2021, with white sand beaches and crystalline waters tainted by toxic smoke plumes and plastic trash pollution. In response, the Maldivian government implemented a ban on open trash burning and single-use plastics later that year. To assess the effectiveness of these policy changes, researchers from Duke University employed advanced AI techniques to analyze satellite images of the islands.
The research utilized machine learning and image segmentation to identify smoke plumes generated by burning trash on Thilafushi, an artificial island used as a dump, and to determine whether the ban on open burning had been successfully enforced. Transfer learning, a method that leverages knowledge from existing neural networks trained on similar tasks, was employed due to the limited availability of environmental data sets. The AI tool achieved an 88% accuracy rate in classifying images and successfully identified the absence of smoke plumes after the burning ban was implemented.
The success of this AI application in the Maldives opens up possibilities for using similar methods to detect and monitor other environmental issues, such as wildfires, power plant emissions, and industrial pollution. By enhancing the AI tool and incorporating a large citizen science effort, it could provide a powerful tool to monitor environmental problems globally, enabling the public to be more informed and hold accountable those responsible for pollution. The research also highlights the positive impact of the Maldives' efforts to regulate environmental problems and protect their natural resources.
