Remote sensing images classification using moment features and attribute profiles

Abstract

Remote sensing is the acquisition of information about an object or phenomenon without making physical contact with the object. Remote sensing is used in numerous fields, including geography, land surveying and most Earth Science disciplines. In supervised classification, all of the feature extraction methods try to increase the accuracy of classification and simultaneously time of computation. At the present work, we use the moments and Attribute Morphology Profiles (APs) to extract texture information from satellite panchromatic images. We use four conventional moments in pattern recognition such as Geometric, Chebyshev, Legendre and Zernike moments and APs to extract features from remote sensing image. An MP is constructed based on the repeated use of openings and closings by reconstruction of a structuring elements (SE) of an increasing size, applied to a scalar image. Then, we use those two set of features together. The well-known support vector machine (SVM) is used for supervised classification. We compare our proposed method with moments and APs. Different criteria such as average accuracy, overall accuracy, κ statistic and computation time are used for assessment of classification performance.