以甘肃省为试验区,利用单时相MODIS数据的光谱信息,使用最大似然法和基于See 5.0数据挖掘的决策树分类方法,进行了分类对比研究。分类结果表明,加入温度-植被角度TVA和温度-植被距离TVD两个指数后,低植被覆盖区的分类效果得到了改善;基于See 5.0数据挖掘的决策树方法能够快速地建立决策树,且能提高较难识别地物类型的分类精度。 Taking Gansu Province as the experimental area, the classification and comparison studies were conducted based on the spectral information of single-phase MODIS data using the maximum likelihood method and the decision tree classification method based on See 5.0 data mining. The classification results show that the classification results of low vegetation coverage are improved after adding two indices of temperature-vegetation angle (TVA) and temperature-vegetation distance (TVD). Decision tree based on See 5.0 data mining can establish decision tree rapidly and Improve classification accuracy of difficult to recognize feature types.