近红外光谱建模法快速测定桔子精油中柠檬烯的含量

Rapid Determination of Limonene in Orange Essential Oil by Near Infrared Spectroscopy Model

采用近红外光谱分析技术获取桔子精油的光谱信息,结合偏最小二乘法(partial least square,PLS)回归分析提取光谱信息中的差异,通过化学计量学将光谱信息差异与对应的柠檬烯测定值关联,建立桔子精油中柠檬烯含量的近红外定量分析模型。模型的校正相关系数RC为0.997 7,校正均方差为0.412;预测相关系数RP为0.998 2,预测均方差为0.396;交互验证相关系数RV为0.994 0,交互验证均方差为0.662;模型的性能指数为94.7。模型近红外光谱(near infrared spectroscopy,NIR)预测值与标准值之间的误差在-0.8%~1.2%之间。模型有良好的预测能力及稳定性。

The spectral information of orange essential oil was obtained by near infrared spectroscopy, and the spectral information was extracted by partial least squares (PLS) regression analysis.The spectral information difference was correlated with the corresponding limonene measurement value by chemometrics, A calibration model of limonene content in orange essential oil by near infrared spectroscopy was established.The corrected correlation coefficient RC of the model was 0.997 7, the corrected mean square error was 0.412;the predicted correlation coefficient RP was 0.998 2, the predicted mean square error was 0.396;the cross-validation correlation coefficient RV was 0.994 0, and the cross-validation mean square error was 0.662;the performance index of the model was 94.7.The error between the model near infrared spectroscopy (NIR) predicted value and the standard value was between -0.8 % and 1.2 %.The model has good predictive power and stability.