In order to explore the relation between arable soil organic matter content and satellite image spectrum, correlation analy-sis was conducted upon soil data from 80 sampling points in Sanming City and Hyperion hyperspectral satellite images. Then charac-teristic soil index and sensitive wavelength were subject to linear regression and stepwise multiple regression analyses. The results showed that soil organic matter content was highly response to Hyperion hyperspectral bands at wavelength of 782. 95-813. 48 nm. Fitting model based on the first derivative of reflectivity was optimal, with R2 being 0.777 and RMSE being 5.31. With R2 between the measured and predicted value of soil organic matter being 0.809, satellite image spectrum is confirmed to be an reliable indicator of soil organic content.%为了探究耕地土壤有机质含量与卫星影像光谱间的关系,确定土壤有机质的光谱特征,构建土壤有机质含量反演模型.利用Hyperion高光谱卫星影像和福建省三明市80个土壤调查样点分析数据,对土壤有机质与光谱指数相关性进行了分析;在提取特征光谱指数的基础上,分别基于敏感波段和特征指数建立线性模型和多元逐步回归模型.结果表明:土壤有机质含量在Hyperion高光谱782.95~813.48 nm波段具有良好的响应能力;反射率的一阶导数所建立的模型拟合效果最优,其R2为0.777,RMSE为5.31,验证模型有机质实测值与预测值的R2为0.809,表明它能够用于区域有机质含量的快速测定.
展开▼
