The Integration of Textural Analysis and Object-Based Method for Oil Spills Detection Using SAR Images (The Persian Gulf)

Oil spill contamination is one of the significant threats for many countries in the world and also it can cause a serious detriment. Nowadays, the various methods and algorithms have been proposed for oil spills contamination detection using SAR imagery, because synthetic aperture radar (SAR) is very important and valuable means to extract oil spills in the marine environment because of its capabilities. In this article, the texture analysis and object- based are used to extract oil spills from SAR images. In the first step of this study, the co-occurrence matrix method is employed to extract textural features of marine SAR image. This method generates eight textural features that contain contrast, dissimilarity, entropy, angular second moment, mean, variance and homogeneity. Within the second step, the results of the textural analysis are integrated and are segmented using object based technique in eCognition software so that every segment shows different information about oil spills and clean water that make it possible to discriminate them of each other. In the final step, the images are categorized and oil spills are separated with high accuracy from SAR images.