YOLO algorithm for shipping waste discharges detection at Swedish waters based on fused satellite images with ship AIS data

Research Project, 2025 – 2026

Waste discharges from ships pose pressures on the marine environment and public health. For example, the discharges from scrubbers, introduced as an alternate solution for shipping companies to comply with the IMO regulation on Sulphur emissions, are now accounting up to 9 percent of emissions of certain cancer-causing polycyclic aromatic hydrocarbon in the Baltic Sea. Maritime authorities are in an urgent need to develop policy related measures to better regulate the waste discharges from ships at sea. Such policy development strongly depends on good understanding/evaluation of how much and where the wastes are discharged into the sea, what are the environmental and health impact from the discharges, and cost/benefits of possible policy measures to handle the discharges, etc. Fortunately, data from Earth Observation satellites, such as satellite radar images, AIS data, etc., provide a unique view to monitor our seas. However, the large dataset collected does not automatically lead to social/industrial benefits, interpretation and understanding of the data poses a lot of demands to the maritime community. In this project, competences from maritime environmental sciences, marine technology, and computer sciences satellite remote sensing will closely cooperate to develop YOLO algorithms (an AI image recognition platform) to automatically detect both historical and real-time waste discharges from scrubbers and tank cleaning of chemical tankers, based on fused satellite sensing radar images and ship traffic AIS data. Spatio-temporal models of those discharges and their contaminants will be established after cross-validation with other modelling methods and marine environment resources.