Vegetation Information Extraction in the South Qinghai Plateau using Phenology and Supervised Classification

In the South Qinghai Plateau there are problems due to the breadth of scope, complex terrain and disparate climates, and it is important to find ways to increase classification accuracy for broad scale vegetation information extraction. In this paper, vegetation information is extracted by supervised classification combined with division of the whole region into major ecological regions, and vegetation phenological knowledge based on NDVI series data and auxiliary information including DEM and other GIS data. Applying this, the classification result is improved with an accuracy of more than 83.3%. The classification is made more accurate by selecting training areas carefully, using vegetation phenological knowledge, visual interpretation and DEM data and taking land-use data into account.