[关键词]
[摘要]
针对目前市场上枸杞掺假现象严重及枸杞品质分析检测繁琐等问题,本试验采用近红外光谱技术对不同产地的枸杞进行光谱分析,对枸杞的各种化学成分进行定量分析。通过偏最小二乘回归法(PLS),其中30个样品为校正集,10个样品为预测集,利用枸杞的近红外光谱数据建立校正模型后,对枸杞的化学值进行预测。用枸杞样品的主成分在空间的分布对全部样品进行检验,去除异常样品,同时采用残余验证方差作为确定主成分数的评价标准。结果表明:枸杞各成分建立的预测模型,其校正集相关系数RC均在0.93以上,交叉检验相关系数RCV均在0.83以上。同时各成分的校正均方差RMSEC值均小于交叉检验均方差RMSEP值,且这两个数值没有明显差异性,预测值与化学值具有良好的相关性。本试验中,枸杞各成分的化学值与预测值都达到了定量标准,可以进行定量分析。
[Key word]
[Abstract]
To solve the problems in sale and identification of Lycium barbarum L, a quantitative anlysis method was developed based on near infrared spectrum for quantitative analysis of chemical composition of Lycium barbarum L Partial least-squares regression (PLS) was introduced for estabulishment of model calibration, in which thirty samples were used for the calibration set and ten samples were used for the prediction set. Then the chemical values of Lycium barbarum L. were calculated. The result showed that using this prediction model of Lycium barbarum L. composition, the calibrating correlation coefficient RC were above 0.93 and cross-check correlation coefficients RCV were above 0.83. The mean square error correction RMSEC values of each component were smaller than those of the cross check all variance RMSEP values. The predicted values had a good correlation with the chemistry values. Both of chemical value and predictive value of Lycium barbarum L reached quantitative standard, indicating that this method can be used for quantitative analysis of chemical composition of Lycium barbarum L .
[中图分类号]
[基金项目]
国家科技支撑项目(2009BAI72B04)