荧光光谱结合模式识别技术鉴别地沟油

Identification of gutter oils based on fluorescence spectrum combined with pattern recognition technology

拟以荧光光谱为基础建立地沟油的识别方法。分别测定了36种（6类）市售普通食用油和5种不同来源地沟油的荧光发射光谱,并利用主成分分析-分组（PCA-Class）方法构建了地沟油的分类模型建立了快速识别地沟油的方法,同时通过正交偏最小二乘-判别分析（OPLS-DA）法分析得到地沟油的特征光谱波段。结果显示该方法的稳定性好,所建模型累积可信度在95%以上,识别率和拒绝率均达到100%。OPLS-DA分析显示450~545 nm波段为地沟油的特征波段。该实验证实荧光发射光谱结合模式识别技术可有效区别地沟油与普通食用油,可为地沟油的鉴别提供一种新的参考方法。

The purpose of the present paper was to establish a model to distinguish gutter oils from edible oils based on fluorescence spectrum.By analyzing the fluorescence spectrum data of 36 common edible oils and 5 gutter oils with principal component analysis-class（PCA-Class）,a rapid classification method for gutter oils was established.Orthogonal partial least squares-discriminant analysis（OPLSDA） method was also applied to explore the characteristic spectral bands of gutter oils.Results showed that the reproducibility of this method was good,cumulative credibility of the model was over 95% and the recognition rate and rejection rate of gutter oils were both 100%.OPLS-DA analysis showed that the range of 450 nm to 545 nm was the characteristic band of gutter oils.Experiments proved that fluorescence emission spectra combined with pattern recognition technology could effectively distinguish gutter oils from common edible oils and thus offer a new feasibility method for gutter oils identification.