傅里叶变换红外光谱结合模式识别法快速鉴别食用油的真伪

Application of Fourier Transform Infrared Spectroscopy Combined with Pattern Recognition Method for Rapid Authentication of Edible Oil

采用傅里叶变换红外光谱(FTIR)结合簇类独立软模式识别技术(SIMCA)建立了真伪食用油的快速鉴别方法.该方法依据FTIR的指纹特性,收集并分析了53个合格食用油和13个伪造食用油的FTIR谱图;通过对谱图取二阶导数和标准化处理,主成分分析(PCA)提取特征变量;采用SIMCA方法分别随机选取43个合格食用油和9个伪食用油样品的FTIR谱图组成训练集,构建得到真伪食用油的SIMCA分类模型.该模型经过剩余10个合格食用油和4个伪食用油的验证,正确识别率达到了100%.说明FTIR结合SIMCA可能成为快速鉴别食用油真伪的一种新方法.

The rapidly analytical method for authenticity of edible oil was established by Fourier transform infrared spectroscopy（FTIR） combined with soft independent modeling of class analogy（SIMCA）.Based on fingerprint characteristics of FTIR,the spectra of 53 qualified edible oils and 13 false edible oils were analyzed.After preprocessing these spectra data with second derivative and normalization,principal component analysis（PCA） was used to extract the characteristic variables in pattern recognition.Then,43 qualified oils and 9 false oils were selected as training set to establish SIMCA classification model.And the model was validated by other 10 qualified oils and 4 false oils as validation set with the correct recognition rate of 100%.The results demonstrated that FTIR combined with chemometrics could be alternatively used to rapidly and simply determine the authenticity of edible oil.