Abstract: Oil spills represented a greater threat to marine ecosystem and their health. From the recent incident in the Gulf of Mexico, the adverse effects of oil spills on the nature are imminent. Synthetic Aperture Radar (SAR), a high resolution remote sensing imagery, can be effectively used for the detection and classification of oil spills. Pollution caused by Oil spills may appear as dark spots in SAR images. However, these images may contain numerous contents which very closely might resemble to oil spill area, resulting in misidentification. The main aim of paper is the development of algorithms to distinguish oil spills from ‘look-alikes. This paper, with the help of two different Artificial Neural Networks (ANN), describes the development of a new approach to SAR oil spill detection. The first ANN identifies the pixels of a SAR image belonging to candidate oil spill features. The second ANN, on the basis of the feature parameters, classifies objects into oil spills and their look-alikes.

Keywords: Artificial Neural Network, Synthetic Aperture Radar (SAR) imagery, Edge detection, Adaptive thresholding, Image segmentation.