A fast segmentation algorithm of SAR oil slick with an improved watershed

Synthetic Aperture Radar (SAR) works by collecting the polarized electromagnetic waves returns from many radar pulses and processing them into a single radar image, which can work all-day and all-weather, and play a significant role in the detection of the oil slick. There´ re speckle noise and the inconsistencies of overall gray level in SAR image, which brings a lot of interference in SAR images´ segmentation. This paper proposed a fast segmentation algorithm with an improved watershed, which has characteristics of segmenting quickly and high accuracy. The method firstly studies on SAR images´ enhancement method with Speckle Reducing Anisotropic Diffusion (SRAD) to suppress speckle noise, and then marks foreground and background with a self-adaptive block threshold methods to eliminate the impact of the inconsistencies of overall gray level, finally obtains a continuous closed boundary of oil slick using watershed. We demonstrate through several examples the effectiveness of the algorithm in splitting SAR oil slick.