Automatic detection of oil spills in the Gulf of Mexico from RADARSAT-2 SAR satellite data

Abstract

This work aimed to design optimization procedures for oil spill footprint automatic detection from synthetic aperture radar (SAR) satellite data. The main motivation of this work is to utilize a genetic algorithm (GA) without involving post-classification image processing tools for oil spill footprint boundary shape optimizations that involve local and global optimizations. The procedures are operated using sequences of RADARSAT-2 SAR ScanSAR Narrow single beam data acquired in the Gulf of Mexico. The study shows that the GA has high performance for oil spill boundary shape automatic optimization and detection. This provides evidence with standard error of 0.12 and non-significant differences with different acquisition dates. The ScanSAR Narrow mode data shows the extremely existing of 90 % of the oil spill footprint compared to the sea surface roughness and look-alikes. It can be said that Scan SAR Narrow mode can monitor oil spill disasters. In conclusion, the GA can be used as an automatic tool for oil spill without involving other post-image processing classification.