Locating Oil Spill in SAR Images Using Wavelets and Region Growing

An algorithm to locate dark areas in the sea that are candidate to be oil slicks by combining region growing approach and multiscale analysis is proposed and provides a better segmentation result that is achieved by a modified regiongrowing approach.

Abstract

This paper presents an algorithm for spots detection in Synthetic Aperture Radar (SAR) images that can be used to support environmental remote monitoring. Monitoring areas with high frequency of oil spillage by accidental or illegal oil discharges can prevent marine damage spreading. But the presence of speckle noise in SAR images limits the visual interpretation of scenes because it obscures the content. Thus, to get reliable data interpretation and quantitative spots measurements, it is recommended to applying speckle filtering schemes. We propose an algorithm to locate dark areas in the sea that are candidate to be oil slicks by combining region growing approach and multiscale analysis. The multiscale analysis employed by the undecimated wavelet smooths the speckle noise in SAR images while enhances edges. The proposed algorithm provides a better segmentation result that is achieved by a modified region growing approach. The algorithms were tested in real SAR images with oil spillages in the sea.