Neural Networks for Oil Spill Detection Using ERS and ENVISAT Imagery

Synthetic Aperture Radar (SAR) images from satellite missions provide a significant support to oil spill detection applications. On the other hands recent studies have demonstrated the potentialities of artificial neural networks for discrimination, starting from SAR imagery, between oil spills and objects which resemble oil spills (called “look-alikes”). The oil spill detection algorithm basically consists of three steps: the identification of dark spots over the sea, the computing of a set of parameters (features) for each dark spot and the classification of the oil spill candidate using a trained neural network, where the network input is a vector containing the values of the features extracted. In this study we report the results obtained by means of a sensitivity analysis and a robustness analysis of the algorithm. This latter includes the introduction of a processability index of the images to be processed.