Improving tropical forest mapping using combination of optical and microwave data of ALOS

In the tropical regions, due to frequent cloud cover, optical remote sensing usually does not have adequate multi temporal high resolution imagery to monitor phenology of forest. This paper seeks to improve mapping of tropical forest by combination of optical and microwave imagery. The study area is located in the southern part of Vietnam. Firstly, ALOS/AVNIR-2 images were used to create a forest map of the study area. Then, ALOS/PALSAR single polarized and dual-polarized images were used to generate a second forest map of the study area. Discrimination of Planted Forest and Natural Forest is one of the most important purposes of this study. The overall accuracy of ALOS/AVNIR-2 classification result is 77.0%, while after combining with ALOS/PALSAR, it is increased up to 88.2%. The accuracy is higher than 90% for main forest classes.