Oil Spill Detection and Mapping Using Sentinel 2 Imagery

Two object-based image analysis methods are developed for detecting oil spills from known natural outflows as well as light oil spill events using Sentinel 2 imagery. The methods are applied to Sentinel 2 images of a known area of natural oil outflow as well as on a Sentinel 2 image of a recent oil spill event along the south coast of Athens, Greece. The preliminary results are considered very successful and consistent, with a high degree of applicability to other Sentinel 2 satellite images. Further testing and fine tuning of the proposed object-based methodology should be carried out using atmospheric correction and ground truth.