中文摘要：

利用中分辨率的TM遥感影像反演湿地土壤水分．探寻湿地土壤水分和非湿地土壤水分提取的不同之处，对湿地监测具有重要意义。建立地表温度（Ts）与归一化植被指数（NDVI）之间的二维特征空间，用1DL编程在特征空间内提取对应的特征点．拟合温度植被干旱指数（TVDI）法需要的干湿边方程，能快速反演出野鸭湖湿地的土壤水分情况。反演结果与实测值的相关系数为0．860，呈极显著相关，均方根误差为0．1042．平均绝对误差为0．0845。结果证明利用温度植被干旱指数（TVDI）法进行野鸭湖湿地这样的小尺度范围的反演土壤水分含量．方法可行．且中分辨率影像的土壤水分反演精度高于低分辨率影像的反演精度．

英文摘要：

The two-dimensional feature space between the surface temperatures （T,）and normalized differential vegetation index （NDVI） was established by using the TM remote sensing images. The dry and wet side equation needed by the method of arid vegetation index was fitted in the feature space with the programming of IDL. The soil moisture condition of Yeyahu wetland was inversed rapidly. The correlation coefficient between inversion results and measured values was 0.860, with root mean square error of 0.1042 and the average absolute error of 0.0845. The results of soil moisture inversion showed that image retrieval accuracy of middle-resolution was better than that of the low-resolution, proving that using temperature vegetation dryness index （TVDI） to invert the wetland soil moisture content on a small scale such as in Wild Duck Lake wetland was feasible.