论文部分内容阅读
基于高光谱遥感技术快速、无损的检测优势,以新疆渭干河-库车河三角洲绿洲为例,探讨利用反射光谱来预测土壤含盐量的可行性。利用野外采集的土壤样本,在实验室内测得了土壤含盐量及原始光谱反射率。利用光谱分析技术计算高光谱指数,与土壤样本含盐量进行相关性分析,筛选出土壤含盐量的光谱特征波段,基于逐步多元线性回归和偏最小二乘回归建立土壤盐分动态监测模型。通过精度检验,结果表明:基于偏最小二乘回归方法,以对数二阶微分光谱特征波段所构建的盐渍化遥感监测模型最优,模型的稳定性和预测精度最高。利用反射光谱来预测土壤含盐量可实现区域尺度上的土壤盐渍化实时监测和评价。
Based on the rapid and non-destructive detection of hyperspectral remote sensing, taking the Weigan-Kuqa River delta oasis in Xinjiang as an example, the feasibility of using soil reflectance spectroscopy to predict soil salinity was discussed. Using the soil samples collected in the field, the soil salt content and the original spectral reflectance were measured in the laboratory. Spectral analysis was used to calculate hyperspectral index and correlation analysis with soil sample salinity. The spectral bands of soil salinity were screened out. The soil salinity dynamic monitoring model was established based on stepwise multiple linear regression and partial least-squares regression. The results of precision test show that the model based on partial least-squares regression method is optimal for the salinization remote sensing monitoring model based on logarithmic second-order differential spectral characteristic band, and the model has the highest stability and prediction accuracy. Real-time monitoring and assessment of soil salinization at the regional scale can be achieved by using reflectance spectroscopy to predict soil salinity.