Abstract/Summary

The increasing complexity of maritime risks and threats requires accurate and timely identification for environmental and human safety. Satellite observa-tions enable comprehensive surveillance of large maritime areas, which is es-sential for detecting and responding to environmental changes and potential threats. The Horizon Europe project EURMARS aims to develop and validate a multi-purpose observation platform to enhance detection capabilities for various risks and threats. This paper introduces a novel Earth Observation (EO) algorithm based on Object-Based Image Analysis (OBIA), employing a You Only Look Once (YOLO) -v9 model to process data from open-access sat-ellites (Sentinel-1, Sentinel-2, Landsat 8, 9) and video from the microsatellite NEMO-HD. Automatic Identification System (AIS) data are used to ensure comprehensive monitoring and validate the method’s results. Satellite im-agery with AIS data integration is a critical element of the vessel tracking methodology, significantly improving the accuracy and reliability of maritime surveillance. Real-life demonstrations have confirmed the method’s effective-ness in enhancing maritime security and facilitating early detection and re-sponse to threats.