Classification of high resolution satellite images
Anders Hans Karlsson · Infoscience (Ecole Polytechnique Fédérale de Lausanne) · 2003
In this thesis the Support Vector Machine (SVM)is applied on classification of high resolution satellite images. Sveral different measures for classification, including texture mesasures, 1st order statistics, and simple contextual information were evaluated. Additionnally, the image was segmented, using an enhanced watershed method, in order to improve the classification accuracy.