Automatic oil slick detection from SAR images: Results and improvements in the framework of the PRIMI pilot project

An automatic system capable of discriminating oil spills from other similar sea surface features in Synthetic Aperture Radar images has been developed and tested. This system, called Oil Spill Automatic Detector (OSAD), was originally conceived for C-band SAR images (mostly ERS PRI) and afterward adapted to ENVISAT data. In the framework of the Progetto pilota Rilevamento Inquinamento Marino da Idrocarburi (PRIMI) national project sponsored by the Italian Space Agency, the OSAD system has been greatly improved and is now able to process L- and X-band images from various satellites as well. OSAD performance, confirmed using a different dataset of verified slicks, shows an a priori overall correct classification of 80%. Moreover, new features have been added, such as an enhanced land masking algorithm, a built-in wind and wave extraction module, and oil spill characterization. OSAD has been integrated into a complex hardware and software architecture for operational sea monitoring, alarm generation, and oil slick drift forecasting. The system?s detection capabilities have been validated during a measurement campaign in the Mediterranean Sea. The new improved system is described herein, with special attention to latest enhancements.