Different approaches have been proposed for detecting and classifying oil spills on SAR data. Several of these are based on training datasets which are used to characterize this phenomenon statistically. In case of images employed for the analysis having different pixel spacing or radiometric resolution to those used in the training set, a new classification template is required. A completely new training dataset and an algorithm optimisation are also needed. In the present paper we present an oil spill detection system which was originally developed for ERS. This has been generalised and put to use for processing ENVISAT data also. Performance of the classification process has been tested using a set of confirmed slicks, which were present on both ERS and ENVISAT images simultaneously. The results are here presented and discussed.
A generalised algorithm for oil spill detection on ers and envisat sar images
TRIVERO, Paolo;BIAMINO, Walter;
2007-01-01
Abstract
Different approaches have been proposed for detecting and classifying oil spills on SAR data. Several of these are based on training datasets which are used to characterize this phenomenon statistically. In case of images employed for the analysis having different pixel spacing or radiometric resolution to those used in the training set, a new classification template is required. A completely new training dataset and an algorithm optimisation are also needed. In the present paper we present an oil spill detection system which was originally developed for ERS. This has been generalised and put to use for processing ENVISAT data also. Performance of the classification process has been tested using a set of confirmed slicks, which were present on both ERS and ENVISAT images simultaneously. The results are here presented and discussed.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
