基于液芯光纤的激光诱导荧光食用油种类鉴别研究

组建了一套基于液芯光纤的激光诱导荧光食用油鉴别装置。研究了不同液芯光纤长度对食用油激光诱导荧光光谱的影响，分析了不同种类食用油激光诱导荧光光谱之间的差异。八种食用油共320份样本荧光数据在1 m长液芯光纤内采集，采用主成分分析方法对食用油荧光数据进行降维处理，利用偏最小二乘判别分析(PLS-DA)方法建立食用油种类的鉴别模型。结果表明，使用液芯光纤后，食用油荧光强度得到较大的增强。随着液芯光纤长度增加，食用油荧光特征峰逐渐增加并且食用油的激光诱导荧光光谱会产生红移现象，当液芯光纤长度超过80 cm后，红移趋于饱和。不同食用油的荧光光谱形状差异较大，可用于区分不同种类食用油。利用主成分1和主成分2绘制的主成分得分图显示，不同种类食用油呈现很好的聚集。当选用主成分数为10时，建立的PLS-DA食用油种类鉴别模型对训练集和预测集样本识别率均达到100%。说明本装置用于食用油种类的快速鉴别具有较高的准确性。

Identification Study of Edible Oil Species with Laser Induced Fluorescence Technology Based on Liquid Core Optical Fiber

A laser-induced fluorescence detection set based on liquid core optical fiber was established in this study.Eight edible oils were discriminated by using this detection set combined with chemometrics method.The effect of length of liquid core optical fiber on laser induced fluorescence spectrum was explored,and the differences between the spectra of different edible oils were analyzed.The fluorescence spectra of 320 samples covering 8 types of edible oil were measured in 1 meter liquid core optical fi-ber.Principal component analysis was used in fluorescence data dimensionality reduction process.Partial least squares discrimi-nant analysis (PLS-DA)method was used to develop the identification model to distinguish edible oil species.The results showed that the oil fluorescence intensity is greatly enhanced when liquid core optical fiber was used.With the increase of liquid core op-tical fiber length,the peaks of laser induced edible oil fluorescence spectra increased and the fluorescence spectra will produce red shift.The red shift tended to a constant value when the fiber length was more than 80 cm.The fluorescence spectra of different edible oils were quite different,its can be used to distinguish different types of edible oil.Principal component scores chart were get using PC1 and PC2 of edible oils fluorescence data which result in a trend of certain gather of same type of edible oil.The recognition rates of PLS-DA model for the calibration set and prediction set were both 100%.The study shows that the devel-oped device in this study has high accuracy for identifying the edible oil species.