中文摘要：

为了研究高光谱遥感在土壤含盐量和pH反演中的作用,于内蒙古河套灌区义长灌域土壤表层（0～5cm）随机采样,并对比分析3种高光谱反演土壤含盐量和pH的方法。结果表明,利用高光谱反演土壤全盐含量和pH是可行的,其中偏最小二乘法建立模型的稳定性最高（R2分别为0.957和0.987）,预测能力最好（RMSE分别为0.682和0.108）,而特征波段法和特征指数法的效果一般。

英文摘要：

In order to study the application of hyperspectral remote sensing in retrieval of soil salinity and pH, topsoil samples in Yichang irrigation district of Inner Mongoliawere collected randomly to analyse the inverse of soil salt content and pH from laboratory＇s hyperspectral remote sensing in three different meth- ods. The results demonstrated that the models based on partial least square method were the most stable （R2 were 0. 957 for salt content and 0. 987 for pH） and had the greatest predictive ability（RMSE were 0. 682, 0. 108 for salt and pH）, while the other two methods were not good. Inverse salinity and pH of soil by means of hyperspectral remote sensing was feasible.