Classification of Remote Sensing Image Based on Multi-parameters of Land Surface
基于多地表特征参数的遥感影像分类研究

The land cover characteristic is the important parameter of land surface information,which is the essential factor to understand the change temporal-spatial information of land surface.In this paper,the method of classification based on multi\|parameters of land surface was proposed.The data of the experimenter was Landsat ETM+ image of Wuhan.The Vegetation Index of Universal Pattern Decomposition (VIUPD),land surface temperature and the character of texture were carried.According to the method of Decision Tree,the image was classified in the consideration of the spectral and spatial information.The results demonstrated that the method of classification based on multi\|parameters of land surface was superior to traditional Decision tree based on the spectral reflectance and Supported Vector Machine (SVM).The accuracy of classification based on multi\|parameters of land surface was higher than that of the other algorithms.