Towards Automatic Recognition of Wakes Generated by Dark Vessels in Sentinel-1 Images

The recognition of wakes generated by dark vessels is a tremendous and interesting challenge in the field maritime surveillance Synthetic Aperture Radar (SAR) images. paper aims at assessing detection performance different scenarios processing Sentinel-1 SAR images along with ground truth data. Results confirm that Radon-based approach an effective technique for wake-based vessels, they lead to deeper understanding effects sea wind conditions. In general, best applicative scenario marine image characterized homogeneous clutter; presence natural surface film or strong transition from low speed areas more windy zones worsen performance. Nonetheless, proposed features vessel capabilities identifying their wakes, without any priori knowledge positions.