中文摘要：

快速获取土壤含盐量是监测、治理土地盐碱化问题的前提条件。以我国重要苏打盐碱土区—白城市为研究对象,以我国的高分一号遥感影像（GF-1）为数据源,结合研究区实地采样的化验数据,定量反演白城市土壤含盐量。首先对遥感影像进行辐射校正、大气校正、图像裁剪及图像镶嵌等预处理;再将影像的反射率及其变换形式与土样含盐量的化验值进行相关性分析,获得盐碱的敏感波段;最后以多元逐步回归分析的方法建立土壤含盐量反演模型,反演研究区土壤含盐量。研究结果表明：GF-1遥感影像具有较高的分辨率,其反射率与土壤含盐量呈显著正相关,将反射率进行适当的数学变换可以提高与含盐量的相关性,以第2波段指数、第4波段倒数、第4波段倒数的对数变换形式建立的反演模型具有较高的精度与稳定性,模型判定系数R2=0.846,均方根误差RMSEcal=0.522。利用GF-1遥感影像反演土壤含盐量是一种快速、稳定、可靠的方法。

英文摘要：

The fast acquisition of soil salinity is the precondition of the monitoring and the management of land salinization. Baicheng, a city located in one of the significant soda saline-alkali soil areas, is used as the object of study and the remote-sensing image from GF-1 as data source while combining the chemical tests data collected from the field sampling of the research area, this paper conducts a quantitative inver- sion of the soil salinity of Baicheng City. Firstly, pretreatments of the radiation correction, atmospheric correction, image cropping and image mosaic, etc. are conducted. Then the correlation analysis of the reflectivity of the image and transformation forms with the laboratory values of the soil sample＇s salinity is conducted so as to get the sensitive band of the saline-alkali. Finally, the inversion model of the soil salinity is es- tablished to conduct the quantitative inversion on the soil salinity of the research area by using the method of multiple stepwise regression a- nalysis. The research results indicate that the remote-sensing image of GF-1 has a relatively higher resolution ratio and that its reflectivity has a significant positive correlation with the soil salinity. Proper mathematical transformations would improve its correlation with salinity. The in- version model established by the second band index, the fourth band reciprocal and the logarithmic transformation forms of the fourth band re- ciprocal has a relatively higher precision and stability with the model determination coefficient R2 = 0.846, root-mean-square error RMSEcal = 0.522. The utilization of the remote-sensing image from GF-1 is a fast, stable and reliable method in conducting the quantitative inversion on soil salinity.