利用ASD FieldSpec Pro高光谱仪在室内条件下测定了风干土壤样品的可见–近红外光谱，分析了不同土壤深度以及不同植被覆盖下土壤光谱反射率曲线形状变化和土壤有机质含量的变化特点，并针对黄土台塬地区的土壤光谱反射率进行一阶导数、倒数的导数、对数的导数、倒数的对数的导数和对数的倒数的导数的变换，并与土壤有机质含量进行多元逐步回归分析，偏最小二乘回归分析(PLSR)，根据特征光谱段建立多元逐步回归模型以及偏最小二乘回归模型，并验证。综合分析得出，偏最小二乘法在土壤有机质含量高光谱预测研究中更具优势。
The visible-near-infrared spectra of air-dried soil samples were determined by ASD FieldSpec Pro hyperspectral spectrometer under indoor conditions. The changes of soil spectral reflectance curve shape and soil organic matter content under different soil depths and different vegetation cover were analyzed, and the loess was analyzed. The spectral reflectance of the soil in the Taihu area is transformed by the first derivative, the derivative of the reciprocal, the derivative of the logarithm, the derivative of the logarithm of the reciprocal and the derivative of the reciprocal of the logarithm, and the multivariate stepwise regression analysis with the soil organic matter content. Least Squares Regression Analysis (PLSR) was established based on the characteristic spectral segments to establish a multivariate stepwise regression model and a partial least squares regression model. The comprehensive analysis shows that the partial least squares method has more advantages in the study of soil organic matter content hyperspectral prediction.