DARTIS
Oil pollution in the oceans is a threat to marine life and coastlines. It is not only caused by maritime accidents but also by illegal discharges of oily waste. The European Sentinel-1 mission enables continuous radar-based monitoring of the oceans.
The DARTIS project has developed an automated oil spill detection and early warning system off the coast of Israel.
DLR’s SAR Oceanography team established the automated detection chain, including Sentinel-1 SAR processing, a custom-trained deep learning-based object detector for oil spills, and segmentation of the identified oil spills into binary masks. To train and validate the object detector, a large dataset of 9768 oil objects in 5930 Sentinel-1 scenes from 2015 to 2018 was collected and manually labelled.
Based on the oil binary masks, our partners simulate the expected oil trajectory and alert the decision makers. It takes around 2–2.5 hours from downloading newly acquired Sentinel-1 data to delivering oil binary masks to the partner institute