A Combined Approach of Retinex & Spatial Kernel Fuzzy C-Means Clustering for Detection of Oil Spills in Satellite Imagery

Abstract

Satellite images published by NASA Earth Observatory mainly affected with low contrast and detecting of oil spillage is quite cumbersome. Here, in this paper a novel approach for detection of oil spills in satellite imagery using Retinex and Spatial Kernel Fuzzy C-Means (SKFCM) Clustering is presented. Firstly the images are enhanced using retinex algorithm which gives more details within an image. Spatial information of enhanced images is been incorporated to kernel fuzzy membership functions for clustering thereby forming different clusters of oil spill and other regions. The results obtained show the profoundness of this approach.