Karst forest type discrimination in southwest China using spaceborne polarimetric SAR data

Karst forest physiognomy occupy a large area of Guizhou, southwest China. It is a rare forest resource in the earth and urgently needed to carry out protection. Due to synthetic aperture radar (SAR) data's ability to acquire images through clouds, it was tested as an alternative to optical data to map changes of land use/land cover, to estimate biophysical parameters of vegetation types, and to detect deforestation. The main goal of this paper is to analyze the potential of the spaceborne full polarimetric data in distinguishing the different types of forest in southwest China. Different polarimetric target decompositions, such as eigenvetor-based decomposition (Cloude-Pottier's decomposition and Touzi Decomposition) and scattering model-based decomposition (Freeman decomposition), were used in this paper to extract forested areas from the scene. To further divide the extracted forested area into deciduous and coniferous forest., Supervised polarimetric classification procedures based on Freeman decomposition is presented in this paper.