A SELF-ORGANIZING-MAP BASED METHOD FOR FINDING PLASTIC PATCHES ON THE SEA

The presence of waste in the oceans, in particular those made of plastic, now represents an emergency on a planetary scale due to the implications on the ecosystem and directly on human health. The plastic dispersed in the environment is collected by sea currents and tends to accumulate, giving rise to formations as large as continents. At the same time, plastic waste tends to break up, giving rise to the phenomenon of micro-plastics. Due to the sea currents, the problem that originates locally necessarily produces global effects, for which nations are urgently called upon to implement coordinated solutions. A first fundamental step to take is to identify effective methods to identify the sources from which this waste is dispersed. This article presents a method of analysis of multispectral satellite images, based on artificial neural networks, which allow to identify the presence of different percentages of plastics in sea water.