Classification of Panchromatic Images Using Ripplet Transfrom and LBP Methods

Abstract— Today, remote sensing is the most effective method is extracting texture features. Image classification is done in two steps: Image feature extraction and automatic classification of these features. In the feature extraction step, RippletI, RippletII, Curvelet and Ridgelet transforms were used. These transforms yield appropriate results in identifying borders and edges of the figures. LBP method is simple got accurate method for identifying index class distribution. Hence, using staking method (combine ripplet transform and LBP methods) results in higher number of features vectors and improves the classification accuracy as much as 5%. SVM classify is used in classification step. Experimental results has been performed two databases (south Tehran and brotza).