Hierarchical classification of SAR data with feature extraction method based on texture features

In this study hierarchical classification structure and feature extraction method based on texture features are applied on SAR data. The most important feature of hierarchical classification is to break down a complex decision-making process into a collection of simpler decisions. In order to achieve more complex analysis it is advantageous to use binary decision trees, in which the decision between only two classes must be assigned at each node. Pixel based feature extraction methods reduce classification performance because of the speckle and also conventional texture analysis is not applicable to every part of an image. Therefore, a decision-making process, which can be applied to every pixel of an image, is required. The results show that computation time and accuracy of classification process are improved.