Marine Safety (Using ML for Tracking Marine Debris)

In this Video from the EUMETSAT MOOC “AI for Earth Monitoring” Dr Lauren Biermann, discusses how satellite technology, specifically Sentinel-2, is used to detect and track marine debris—such as driftwood and man-made waste—that poses navigational hazards. The satellite’s spectral capabilities allow it to distinguish between water and floating objects, making it valuable after events like floods, when large amounts of debris enter the ocean.

By applying machine learning—particularly Naive Bayes classifiers—different types of debris such as plastic, seaweed, and driftwood, can be classified based on probability and a carefully constructed training dataset. This method significantly accelerates the analysis process and makes large-scale, even global, monitoring of marine debris feasible—something that would be unmanageable through manual efforts alone.