Chapter 11 - Automatic detection of oil spills from SAR satellite data using genetic algorithm

Abstract

This chapter presents an approach for automatic detection of oil spills in SAR data. The method is based on the genetic algorithm. This chapter shows that the RADARSAT-2 data prove the potential to identify even small oil spill pixels from the surrounding large oil spill pixels by using the optimization tool of genetic algorithm. In fact, the crossover process begins with a cell of small oil spill pixels dominated by small backscatter. In the cancellation step, the isolated oil spill starts to be generated from the surrounding sea. This confirms the terms of the receiver–operating characteristic (ROC) area; small oil spill pixels exhibit an area difference of 30% compared with 60% for large oil spill pixels, 3% for look-alikes, 7% for sea roughness, and a P-value of less than 0.0005.