Oil spill monitoring on water surfaces by radar L, C and X band sar imagery: a comparison of relevant characteristics

During last years, several studies related to remote sensing technologies analyzed the processes to extract and classify slicks from SAR imagery. These images are used, among other purposes, for monitoring coastal and marine waters pollution where oil floating on the surface becomes visible because it damps the short gravity-capillary waves that are responsible for the radar backscattering [14]. Nowadays an important number of SAR images are available and this number will increase in coming years thanks the launch of Cosmo-Skymed 2nd generation, recent availability of Sentinel-1, ALOS Palsar-2 products and future SAOCOM launch. That will provide information suitable to support decision makers in managing emergencies or potential disasters. The present study show the results obtained from 190 regions of interest extracted from a set of X, C and L Band images, where a database related to spatial, textural, spectral and contextual characteristics of the features detected was ingested into a neural network algorithm. The classification process reached percentages of up to 95% of cases of oil spills and look-alikes correctly classified depending on the wavelength, the polarization and incidence angle.