Model-Based Oil Slick Thickness Estimation Using Artificial Neural Network

In this study, the Artificial Neural Network technique is used to retrieve quantitative parameters of marine oil spill on Synthetic Aperture Radar imagery. In fact, while Synthetic Aperture Radar has been widely exploited to obtain morphological features of sea oil spills as extent and shape, its potential to extract oil thickness information has been underexplored. Hence, an artificial neural network is proposed that have been trained and tested using the damping ratio predicted by a microwave scattering model consisting of Advanced Integral Equation Method in combination with the local balance damping model and a layered-medium dielectric model. Experiments are performed on a L-band Uninhabited Aerial Vehicle Synthetic Aperture Radar image collected during the DeepWater Horizon oil spill accident. The inversion results show that the central area of the slick is the thicker part of the oil emulsion (2 – 4 mm thickness), surrounded by thinner oil film whose thickness is lower than 1 mm.