科尔沁沙地土壤表层水分遥感反演模型研究

Research on the soil surface moisture retrieval model by remote sensing in Horqin sandy

为了对三北防护林体系和科尔沁沙地的沙漠化进行监测,选择内蒙古翁牛特旗作为实验区,利用LandsatTM遥感数据对实验区进行了地表土壤水分反演。文中采用热惯量法、温度植被干旱指数法和BP神经网络法进行地表土壤水分的反演,3种算法均能较好地反映地表水分的空间分布趋势。并对三种反演算法建立的土壤含水量遥感信息模型进行了精度检验。统计结果表明:BP神经网络法建立的遥感信息模型精度最高;温度植被干旱指数法建立的遥感信息模型精度居中;基于热惯量法建立的土壤含水量遥感信息模型精度最低。实验结果为三北防护林和科尔沁沙地沙漠化监测奠定了部分数据基础。

In order to monitor the desertification of the three-north shelterbelt and Horqin sandy,the paper selected Wengniute county of Inner Mongolia as an experimental area and displaied the surface soil moisture by using Landsat TM remote sensing data.This paper adopted the thermal inertia method,temperature vegetation drought index method and the BP neural network method to display the surface soil moisture,all of the three algorithms can better reflect the spatial distribution trend of surface soil moisture.And respectively executed precision inspection.Statistical results showed that the BP neural network method possesed the highest precision,temperature vegetation drought index method was secondary,the thermal inertia method was lowest.The experimental results can supply the part data foundation for the monitoring in three north shelterbelt and the desertification of Horqin sandy land.