Mapping tropical forest vegetation from Landsat TM based on knowledge discovering

This paper focuses on mapping the tropical forest vegetation by using Landsat TM and ancillary data such as DEM in Xishuangbanna autonomous prefecture, Yunnan Province, Southwest China. The spectral knowledge of the tropical forest vegetation was obtained by image analysis and field work. The relationship knowledge of the different tropical forest vegetation types with the ancillary data was obtained. A classification method based on knowledge was presented and applied to classify the entire image. It shows that more forest vegetation types with high accuracy can be obtained by the method.