An efficient method to solve the classification problem for remote sensing image

Remote sensing image classification is different to identify the best classification model due to lack of a suitable classification method. Most traditional approaches only focused on using the spectral or spatial information to train classification model. However, these methods may ignore the related information of the image-itself. Remotely data not only a mere collection of independent and identically distributed pixels. Therefore, an efficient classification method is introduced in this paper. The proposed method deals with the information provided by the remote sensing image. Based on the idea of fisher linear discriminant analysis (FLDA), a definition of the same areas and different areas are considered in images, the information of same areas associated with each pixel is modeled as the within-class set, and the information of different areas associated with mean pixel of each same areas is modeled as the between-class set. Therefore, a projection matrix (PM) can be obtained by using within-class and between-class sets with the help of FLDA criterion. Then the PM is jointly used for the classification through a support vector machine (SVM) or K-nearest neighbor (KNN) classifiers formulation. Experiments on two remote sensing images are performed to test and evaluate the effectiveness of the proposed method.