基于高光谱的土壤有机碳含量预测研究

Prediction Research of Soil Organic Carbon Content Based on Hypers Pectral

[目的]对土壤有机碳含量进行预测研究。[方法]利用高光谱仪对表层土壤进行光谱测定并且进行光谱数据的预处理,通过多元线性逐步回归(SMLR)和偏最小二乘回归(PLSR)方法对土壤有机碳含量进行预测,并对2种模型的精度进行比较。[结果]LSR模型的精度高于SMLR模型。[结论]偏最小二乘回归法优于多元逐步回归法,对有机碳的预测具有更好的效果。

[ Objective ] To predict soil organic carbon content. [ Method ] Surface soil was detected by high spectrometer spectrometric and spectral data was treated, through stepwise multiple linear regression （ SMLR） and partial least-squares regression （ PLSR） method, soil organic carbon content was predicted ,and the accuracy of the two models was compared. [Result]The accuracy of PLSR model was higher than SMLR model.[Conclusion ] PLSR method is better than SMLR method in forecasting organic carbon.