作者: Attilio Gambardella , Giorgio Giacinto , Maurizio Migliaccio
DOI: 10.1109/IGARSS.2008.4779618
关键词: Synthetic aperture radar 、 Computer science 、 Remote sensing 、 Artificial neural network 、 One-class classification 、 Classifier (UML) 、 Feature extraction 、 Feature selection 、 Oil spill 、 Data mining 、 Contextual image classification
摘要: A novel approach to oil-spill classification, based on the paradigm of one-class is proposed. Basically, a classifier trained using only examples oil-spills, instead oil-spills and look-alikes, as in two-class approaches. In addition, large number candidate features have been considered literature, feature selection algorithm, objectively select most effective subset, Results two case study datasets are reported validate proposed approach.