Adapting Neuro Fuzzy System for Oil Spills Detection From Sea Sar Images

Marine oil slicks represent both an environmental and Economical concern since they seriously effect fragile marine and coastal ecosystem. Accurate detection of oil slicks can help identification and management of polluted areas and assessment of oil slicks drift in order to protect the coasts. Today, Synthetic Aperture Radar (SAR) satellites are the main data sources to detect the oil spills discharged into the sea with sufficient accuracies. In this study, we present SAR image classification process to extract dark spots using neuro-fuzzy systems. Thus, we first performed multi-scale analysis image to extract textural descriptors that can characterize the surface roughness. For this purpose, we used the distance measure similarity of kullback-leibler (KLD). Then, we integrated these textual descriptors into Adaptive Neuro-Fuzzy Inference System (ANFIS) classifier to extract dark spots. The proposed process was tested on SAR image covering Algerian coasts and gave good spots detection with minimum false alarms.