Mapping tropical forest vegetation from Landsat TM images based on fusion of knowledge and geo-data

Tropical forests play a crucial role in the function of our planet and in the maintenance of life. Tropical forest vegetation maps are very important for managing tropical forests. Mapping tropical forest vegetation only by spectral-based remote sensing techniques has proven to be problematic. The objective of the study is to develop a rule-based model to identify different forest types using Landsat TM images and GIS. In this paper, we developed the rule-based model to identify different forest types in Xishuangbanna, P.R. of China, using two temporal Landsat TM images and geo-data such as DEM, rainfall and temperature. The results show that the method put forward is useful and effective in tropical forest vegetation mapping, which can effectively integrate multi-knowledge and multi-resource data to identify the tropical forest vegetation types with higher accuracy.