Spectral and textural information of multisensor data for land use classification in metropolitan area

Abstract

This paper presents the potential of RADARSAT and Landsat-TM images to discriminate land use categories in metropolitan area of Jakarta, Indonesia. Spectral information used in this study, is standard principal component analysis of Landsat-TM. Four of the most common first and second order texture statistics found in the literature were used. They are variance, entropy, angular second moment, and contrast. Several datasets were generated using spectral extraction, textural features, and their combination. Based on the maximum likelihood classifier, land use categories of the study area were discriminated. The overall accuracies and the kappa statistics were analyzed and compared. The result shows that combined use of spectral and texture information together significantly improved the accuracy of land use classification.