基于NIR结合化学计量法牛肉糜掺假检测  

Detection of beef surimi adulteration based on NIR and chemometrics

作　　者：李锋[1] 冷拓 唐丽君 甘蓓 陈奕 LI Feng;LENG Tuo;TANG Li Jun;GAN Bei;CHEN Yi(School of Food,Nanchang University,Nanchang 330031,China;Jiangxi Institute of Food Inspection and Testing,Nanchang 330001,China;Jiangxi Product Quality Supervision and Testing Institute,Nanchang 330052,China)

机构地区：[1]南昌大学食品学院,江西南昌330031 [2]江西省食品检验检测研究院,江西南昌330001 [3]江西省产品质量监督检测院,江西南昌330052

出　　处：《南昌大学学报（理科版）》2022年第5期557-562,共6页Journal of Nanchang University(Natural Science)

基　　金：国家重点研发计划项目(SQ2019YFE010931);国家自然科学基金资助项目(21866021)。

摘　　要：采用近红外(Near Infrared,NIR)技术(12500~5400cm^(-1))快速无损检测牛肉糜的掺假。判别分析(Discriminant Analysis,DA)、主成分回归(Principle Component Regression,PCR)等功能强大的化学计量学技术被用于掺假检测和掺假水平预测模型。通过选择适当的光谱波长和使用不同的光谱预处理方法,优化了DA和PCR模型。选择特定波长和使用(无预处理方法)的DA模型分类率达到了100%。基于全波长和使用(无预处理方法的)PCR的最佳预测模型的相关系数Rp为92.21%,样本的预测均方根误差(Root Mean Square Error Of Prediction,RMSEP)为9.80。研究结果说明NIR技术对牛肉糜的掺假体系适用。The near infrared(NIR)technology(12500~5400 cm^(-1))was used to detect the adulteration of beef surimi.Discriminant analysis(DA),principal component regression(PCR)and other powerful chemometrics techniques were utilized in adulteration detection and adulteration level prediction model.The DA and PCR models were optimized by selecting appropriate spectral wavelength and using different spectral pretreatment methods.The classification rate of DA model with specific wavelength and without pretreatment method reached 100%.The correlation coefficient(RP)of the best prediction model based on full wavelength PCR and PCR without pretreatment method was 92.21%,and the root mean square error of prediction(RMSEP)was 9.80.The results showed that NIR technology is suitable for the adulteration system of beef surimi.

关 键 词：掺假检测 牛肉糜 近红外光谱 化学计量法 

分 类 号：O65[理学—分析化学]