基于三维荧光光谱的枸杞子储藏年份鉴别研究

Identification of Lycii Fructus Storage Years Based on Excitation-emission Matrix Fluo-rescence Spectroscopy

目的利用三维荧光光谱结合模式识别技术构建枸杞子储藏年份的鉴别方法。方法采集枸杞子提取液的三维荧光光谱数据,分析当季及不同储藏年份(1年、2年和3年)枸杞子样品的三维荧光光谱特征差异,利用主成分分析结合线性判别分析(PCA-LDA)构建枸杞子不同储藏年份的分类模型。结果不同储藏年份枸杞子提取液的三维荧光光谱特征相似,均具有5个较明显的荧光峰,且荧光峰位置和峰形基本一致,但荧光强度差异显著。当PCA-LDA模型的主成分数选择为5时,其训练集、测试集及预测集分类准确率均达到100%,且灵敏度和特异性也均达到100%。结论基于三维荧光光谱数据利用PCA-LDA模型可以快速准确鉴定不同储藏年份枸杞子样品。

Objective To establish a method for identifying the storage years of Lycii Fructus based on excitation-emission ma-trix fluorescence spectroscopy combined with pattern recognition technology.Methods The excitation-emission matrix fluores-cence spectra of the extract of Lycii Fructus were measured by fluorescence spectrophotometer.And the differences in excitation-emission matrix fluorescence spectra characteristics of Lycii Fructus in the current season,stored for one year,stored for two years,and stored for three years were analyzed.Then,principal component analysis-linear discriminant analysis(PCA-LDA)was used to construct a classification model of Lycii Fructus with different storage years.Results The excitation-emission matrix fluorescence spectra characteristics of Lycii Fructus extracts of the four storage years were similar,with 5 obvious fluorescence peaks,and the positions and shapes of the fluorescence peaks were basically the same,but the fluorescence intensity was signifi-cantly different.When the number of principal components of the PCA-LDA model was selected to be 5,the correct classifica-tion rate of the training set,test set and prediction set all reaced 100%,and the sensitivity and specificity of all classes also reached 100%.Conclusion The establishment of a PCA-LDA model based on excitation-emission matrix fluorescence spectral data could quickly and accurately identify samples of Lycii Fructus in different storage years.