摘要
采用近红外光谱分析技术对麦冬有效成分进行定量分析。对麦冬的原始漫反射光谱采用了Savitzky-Golay平滑、多元散射校正、一阶导数及二阶导数等多种光谱预处理方法进行处理,并结合偏最小二乘法对麦冬总皂苷和总黄酮含量进行了定标建模分析。结果表明,光谱均在经过Savitzky-Golay平滑+多元散射校正(MSC)+一阶导数预处理后得到的定标模型效果最佳,分析结果精度较高,总皂苷模型R^2、RMSEC、Rcv^2、RMSECV、主因子数分别为0.938、0.76、0.901、0.98和6,总黄酮模型R^2、RMSEC、Rcv^2、RMSECV、主因子数分别为0.872、1.04、0.857、1.32和4。模型经过验证样品集检验,预测决定系数达到0.930和0.655。
The active constituents in Ophiogon japonic us（Thunb.）Ker.Gawl were quantitatively analyzed by near infrared（NIR） spectroscopy.The various spectrum pretreatment methods,such as first derivative,second derivative,Savitzky-Golay smoothing and Multiple Scatter Correction（MSC） and so on,were used to deal with the original diffuse light spectrum of Ophiogon japonic us（Thunb.） Ker.Gawl,and the contents of the total flavonoids and total saponins in Ophiogon japonicus（Thunb.） Ker.Gawl were performed to establish calibration modeling analysis.The optimal effect of calibration model was obtained via spectrum treatment of Savitzky-Golay smoothing ＋ Multiplicative Scatter Correction（MSC） ＋ first derivative,and the analytical results had high precisions,and R^2,RMSEC, Rcv^2,RMSECV and factors number of the total saponins were 0.938,0.76,0.901,0.98 and 6, respectively,while those of the total flavonoids were 0.872,1.04,0.857,1.32 and 4,respectively.The model was tested and verified by sample set inspection,and prediction coefficients of determination were 0.930.and 0.655,respectively.
出处
《光谱实验室》
CAS
CSCD
2012年第1期551-555,共5页
Chinese Journal of Spectroscopy Laboratory
基金
国家"十一五"科技支撑计划子课题(No.2006BAI09B05)
关键词
麦冬
皂苷
高异黄酮
近红外光谱
Ophiogon japonicus（Thunb.） Ker.Gawl
Saponins
Homoisoflavonoids
NIR Spectroscopy
