基于6 mm检测皿近红外测量光谱iPLS模型识别复原驼乳参伪水解动物蛋白的研究  

Identification and Restoration of Pseudo-Hydrolyzed Animal Protein of Lacteus Camelus Based on iPLS Model of Near-Infrared Measurement Spectrum of 6 mm Detection Plate

作　　者：苑柯岩 王嵘 王翔翔 薛莉娉 余丽 YUAN Ke-yan;WANG Rong;WANG Xiang-xiang;XUE Li-ping;YU Li(Huhhot City Inspection and Testing Center,Huhhot 010018,China;School of Environment and Energy Engineering,Anhui Jianzhu University,Hefei 230601,China)

机构地区：[1]呼和浩特市检验检测中心,内蒙古呼和浩特010018 [2]安徽建筑大学环境与能源工程学院,安徽合肥230601

出　　处：《光谱学与光谱分析》2022年第10期3143-3147,共5页Spectroscopy and Spectral Analysis

基　　金：国家自然科学基金项目(41601203,51978003);安徽省自然科学基金项目(2108085ME186,1908085QE241,KJ2020A0467)资助。

摘　　要：驼乳因具较高的营养价值和独特的保健作用,已逐渐成为广大消费者所信赖的保健乳制品。由于驼乳产量小而其市场价值高,为在驼乳中参杂参伪提供了利润上的操作空间。随着国家对乳制品中非法添加三聚氰胺的严打态势进一步加强,劣质水解动物蛋白以其高蛋白含量、价格低廉和非法添加隐蔽性强等特点逐渐成为乳制品中参伪的新宠。防范和打击驼乳中参伪劣质水解动物蛋白成为了消费者和驼乳产业从业人员共同面临的巨大挑战,如何快速、便捷、低成本检测驼乳当中参伪动物水解蛋白成为亟待解决的问题。随着近红外光谱分析技术的飞速发展,其分析速度快、成本低、样品无需前处理、操作简便等优点逐渐在石油化工、食品、农业、医药等诸多领域得到广泛应用。采用6 mm测样皿的近红外光谱仪对驼乳参伪不同含量的动物水解蛋白进行测量获得原始光谱矩阵,采用一阶导数、SNV、SG卷积平滑法、一阶导数+SG平滑法、一阶导数+SNV、SG+SNV等方法对原始光谱进行预处理,以全光谱10个主成分回归模型为评价,通过比较原始光谱在不做任何预处理为本体系建立iPLS模型最佳光谱。通过调整主成分计算规模,确定了最佳主成分计算规模为10个。通过调整区间划分数量,以其对应的模型的R^(2)和RMSECV值为评价标准,最终确定了最佳区间划分数量为30。通过实验和计算,在7887.87~7590.87 cm区间得到了主成分数6,相关系数0.9451,RMSECV值0.2001为驼乳掺伪水解动物蛋白最佳预测模型。经内部交互验证,该模型在本体系可以很好地预测复原驼乳中掺伪水解动物蛋白的情况,为相关领域研究提供技术参考。Camel milk has gradually become a health care dairy product trusted by consumers because of its high nutrition and unique health care effects.However,due to the small output of camel milk and its high market value,this provides a profitable operating space for the hybridization of camel milk.With the further strengthening of the state’s crackdown on the illegal addition of melamine in dairy products,inferior hydrolyzed animal protein has gradually become a new favorite for counterfeiting in dairy products due to its high protein content,and low price and strong concealment of illegal addition.Preventing and cracking down on fake and inferior hydrolyzed animal protein in camel milk has become a huge challenge faced by consumers and practitioners in the camel milk industry.How to detect fake and low-cost animal hydrolyzed protein in camel milk has become an urgent problem to be developed.With the rapid development of near-infrared spectral analysis technology in the past ten years,near-infrared spectral analysis technology has gradually become widely used in many fields such as petrochemical,food,agriculture,medicine,etc.widely used.In this paper,the near-infrared spectrometer with a 6 mm sample dish was used to measure the animal hydrolyzed protein of camel milk ginseng with different contents to obtain the original spectral matrix.The original spectra were preprocessed by order derivative+SNV,SG+SNV and other methods,and the 10 principal component regression models of the global spectrum were used for evaluation.By adjusting the calculation scale of principal components,the optimal calculation scale of principal components is determined to be 10.By adjusting the number of interval divisions and using the R^(2)and RMSECV values of the corresponding model as evaluation criteria,the optimal number of interval divisions is finally determined to be 30.Through experiments and calculations,the principal component score of 6 was obtained in the range of 7887.87~7590.87 cm,the correlation coefficient was 0.9451,and the

关 键 词：近红外光谱 间隔偏最小二乘法 复原驼乳 水解动物蛋白 

分 类 号：O657.33[理学—分析化学]