Decision Tree Method Based on NDVI/NDWI/ DEM for land use classification of ALOS Image in Dongguan City

Using the ALOS image of Dongguan City in 2008 as data source, We first selected 8 objective land use types and used a maximum likelihood method for land use classification. We found that the classification accuracy is too low (80%). The main reason is that the ALOS data has few effective bands, and there are too many vegetation and waters in the study area, so it is difficult to distinguish many land use types. To solve the problem, this study combined the characteristics of topography, imported normalized difference vegetation index (NDVI), normalized difference water index (NDWI) and digital elevation model (DEM) data, used decision tree methods for land use classification. The classification accuracy has improved greatly ( 90%). The study showed that in sub-tropical regions in southern China, the modified decision method based on vegetation, water index and DEM is a very useful land use classification method for ALOS data.