The Applications of High-resolution Satellite Images for Land Cover monitoring in Nanji Islands

Abstract

This paper describes the monitoring of land covers using remote sensing technique. IKONOS imagery which is the world's first 1-meter commercialized remote sensing satellite imagery is used to obtain the land cover and its dynamic information. The methods for geometrical correction method, image fusion method and supervised classification method are discussed. The information of the land cover types of Nanji Islands is derived from these images using supervised classification method, threshold method, vegetation index method and man-machine exchange method. The results show that the natural land cover types such as meadows and shrubbery are the main types which occupy about 70% area of the Islands, that the accuracy of the supervised classification method is about 76.31% and the Kappa coefficient about 0.71. Paying the especial attention to the main island, its area and shoreline derived from IKONOS images are 752 sq. km and 34.64 km, respectively. Comparing the statistics values, their biases are 1.46% and 456%.