Classification of land cover in upstream of Lijiang river basin based on decision tree technologies
Aiming at the problem that it is difficult for Land use/cover classification with multispectral data in subtropical mountainous areas due to the complex terrain,various vegetation types,and hard validation of data acquisition,the paper discussed to classify the subtropical vegetation in the mountains by using phenologic information:algorithms of Normalized Difference Vegetation Index (NDVI),Ratio Vegetation Index (RVI),and Normalized Difference Water Index (NDWI)were integrated,and a Decision Tree model was built considering the influence of altitude on vegetation type distribution,which used the phenological and spectral characteristics of different land cover types to classify the upstream of Lijiang river basin into eight land cover types with multi temporal Landsat TM imagery.Result showed that compared with existed methods,the classification accuracy would be improved higher with an overall accuracy of 86.40％ and Kappa coefficient 0.83.
upstream of Lijiang river basindecision treephenologyland use/cover
NETL
NSTL
万方数据
维普
