基于植被分区的全国植被类型分类方法

he high precision land cover classification products has a very important significance for the study of quantitative remote sensing and remote sensing applications. Now there has a lot of free and global coverage land classification products, which is mostly developed by foreign research institutions and personnel. Due to the China regional complex terrain, differences in vegetation structures and in crop planting structures have not been fully studied, the classification accuracy of these products in the area of China is very low, especially the vegetation classification accuracy. Therefore, it is necessary to produce a vegetation type classification product of China region. So according to Chinese regional topography, soil and other information, and based on existing vegetation regionalization, we developed a method of Chinese vegetation classification based on vegetation partition. This method is based on long time series, which can capture the surface information changed with time with the high time resolution, and can improve the classification accuracy by the differences in the time dimension of the object. And we complete the 2012 national land cover classification by using the method. At last, the method of stratified random sampling was carried out to assess the accuracy of the classification results, found that the overall accuracy and Kappa coefficient of the classification products is greatly improved, which the overall accuracy was 90.78%, Kappa coefficient was 0.86.And the product was compared with the MODIS land cover data products, found that the accuracy of this product increased by 61.38% in vegetation than the MODIS land cover data product.