基于近红外光谱技术构建牛羊肉掺假鉴别模型
梁静 , 郝生燕 , 赵祥民 , 李璐璐 , 康景 , 年芳 , 张兆杰 , 唐德富
甘肃农业大学学报 ›› 2023, Vol. 58 ›› Issue (01) : 19 -29.
基于近红外光谱技术构建牛羊肉掺假鉴别模型
Construction of an adulteration identification model for beef and mutton based on near-infrared spectroscopy
目的 为应用近红外光谱技术结合正交偏最小二乘判别分析(orthogonal partial least-squares discrimination analysis,OPLS-DA)方法对牛羊肉掺杂鸡肉和鸭肉的定性鉴别。 方法 采集新鲜的牛里脊肉、羊后腿肉、鸡胸肉、鸭脯肉各41份,用绞肉机绞碎成肉糜,按照0%、25%、50%、75%、100%的掺假比例分别将鸡肉糜和鸭肉糜掺入到牛肉糜和羊肉糜中,得到4种纯肉样品各41份,牛肉掺鸭肉、牛肉掺鸡肉、羊肉掺鸭肉、羊肉掺鸡肉样品各123份,并采集近红外光谱。利用二阶导数(second derivative,2nd Der)、savitzky-golay滤波平滑(savitzky-golay,S-G)、多元散射校正(multiplicative scatter correction,MSC)、标准正态变量校正(standard normal variate correction,SNV)、2nd Der+S-G、2nd Der+MSC、2nd Der+SNV、S-G+MSC、S-G+SNV等方法对原始光谱数据预处理,结合OPLS-DA分别构建牛羊肉掺假定性鉴别模型。根据模型训练集的最高正确判别率以及比较受试者工作特征曲线(receiver operating characteristic curve,ROC曲线)的面积(area under curve,AUC)评价模型的预测性能。 结果 原光谱建立的牛肉掺鸭肉模型和牛肉掺鸡肉模型对训练集鉴别准确率均为97.42%,对预测集的鉴别准确率分别为96%和92%;经S-G+MSC预处理的牛肉掺假模型对训练集和预测集的鉴别准确率分别为97.24%和90%;经SNV预处理的羊肉掺鸭肉模型对训练集和预测集的鉴别准确率分别为96.13%和92%;经SNV预处理的羊肉掺鸡肉模型对训练集和预测集的鉴别准确率分别为96.77%和94%;经S-G+SNV预处理的羊肉掺假模型对训练集和预测集的鉴别准确率分别为97.24%和97.14%。 结论 研究结果为近红外技术结合OPLS-DA方法对定性鉴别掺假牛羊肉提供了参考。
Objective For the qualitative identification of beef and mutton adulterated chicken and duck by using near-infrared spectroscopy technology combined with orthogonal partial least squares discriminant analysis method. Method 41 fresh beef tenderloin,mutton ham,chicken breast and duck breast were collected and minced using a mincer.The chicken and duck meat were mixed with beef and lamb meat at adulteration ratios of 0%,25%,50%,75% and 100%,respectively.The NIR spectra of pure beef,mutton,duck and chicken meat (41 portions for each),beef mixed with duck,beef mixed with chicken,mutton mixed with duck and mutton mixed with chicken (123 portions for each) were recorded.The second derivative (2ndDer),Savitzky-Golay filter smoothing (Savitzky-Golay,SG),multiplicative scatter correction (MSC),standard normal variate correction (SNV),2ndDer+SG,2ndDer+MSC,2ndDer+SNV,S-G+MSC,S-G+SNV and other methods were used to pre-process the original spectral data and combined with OPLS-DA to build a qualitative identification model of beef and mutton adulteration.On the other hand,the predictive performance of the model was evaluated based on the highest correct discrimination rate of the model training set and the area under the receiver operating characteristic curve (ROC) (AUC). Result The model of beef adulteration by Savitzky-Golay combined with multivariate scatter correction processing,the accuracy of the training set was 97.42%; the accuracy of the prediction set was 96% and 92%; the accuracy of the training set and the prediction set was 97.24% and 90%,respectively.24% and 90% respectively,and it was identified over the model of beef adulteration by Savitzky-Golay combined with multivariate scatter correction processing; the accuracy of the training set and prediction set were 96.13% and 92% respectively,and it was identified over the model of mutton adulterated with duck by standard normal variable processing; the accuracy of the training set and prediction set were 96.77% and 94%,13% and 92% respectively,and it was identified on the model of mutton adulterated with duck by standard normal variable processing; the accuracy of the training set and prediction set were 96.77% and 94% respectively,and it was identified on the model of mutton adulterated with chicken by standard normal variable processing; the accuracy of the training set and prediction set were 97.24% and 97.17% respectively,and it was identified on the model of mutton adulteration by Savitzky-Golay combined with standard normal variable processing. Conclusion The results provide a reference for the qualitative identification of adulterated beef and mutton using near-infrared technology combined with the OPLS-DA method.
牛羊肉掺假 / 近红外光谱 / 定性鉴别 / 正交偏最小二乘判别分析(OPLS-DA)
beef and mutton adulteration / near infrared spectroscopy / qualitative identification / orthogonal partial least-squares discrimination analysis (OPLS-DA)
| [1] |
|
| [2] |
|
| [3] |
石长波,姚恒喆,袁惠萍, |
| [4] |
|
| [5] |
|
| [6] |
庄齐斌,郑晓春,杨德勇, |
| [7] |
王慧,何鸿举,张海曼, |
| [8] |
孟一,张玉华,许丽丹, |
| [9] |
郝冬梅,周亚男,王玉, |
| [10] |
於海明,徐佳琪,刘浩鲁, |
| [11] |
|
| [12] |
|
| [13] |
杨清华,彭彦昆,郑晓春.基于阈值选取高光谱感兴趣区域的牛肉掺假检测[J].食品安全质量检测学报,2018,9(11):2710-2715. |
| [14] |
孙宗保,王天真,李君奎, |
| [15] |
朱亚东,何鸿举,王魏, |
| [16] |
|
| [17] |
|
| [18] |
刘友华,白亚斌,邱祝福, |
| [19] |
张玉华,孟一,姜沛宏, |
| [20] |
张丽华,郝莉花,李顺峰, |
| [21] |
|
| [22] |
蒋祎丽,吴晓宗,郝莉花, |
| [23] |
|
| [24] |
石吉勇,李文亭,邹小波, |
| [25] |
|
| [26] |
Weng S Z,Guo B Q,Tang P P,et al.Rapid detection of adulteration of minced beef using Vis/NIR reflectance spectroscopy with multivariate methods [J].Spectrochimica Acta Part a-Molecular and Biomolecular Spectroscopy,2020,230:118005. |
| [27] |
Deniz E,Altuntas E G,Ayhan B,et al.Differentiation of beef mixtures adulterated with chicken or turkey meat using FTIR spectroscopy [J].Journal of Food Processing and Preservation,2018,42(10):e13767. |
| [28] |
白亚斌,刘友华,丁崇毅,基于高光谱技术的牛肉-猪肉掺假检测[J].海南师范大学学报(自然科学版),2015,28(3):270-273. |
| [29] |
白京,李家鹏,邹昊,近红外光谱定性定量检测牛肉汉堡饼中猪肉掺假 [J].食品科学,2019,40(8):287-292. |
| [30] |
白京,李家鹏,邹昊,近红外特征光谱定量检测羊肉卷中猪肉掺假比例 [J].食品科学,2019,40(2):287-292. |
| [31] |
Leng T,Li F,Xiong L A,et al.Quantitative detection of binary and ternary adulteration of minced beef meat with pork and duck meat by NIR combined with chemometrics [J].Food Control,2020,113:107203. |
| [32] |
郭旭生,尚占环,方向文,近红外光谱技术(NIRS)在反刍动物营养研究中的应用现状 [J].光谱学与光谱分析,2009,29(3):641-646. |
| [33] |
张丽华,相启森,李顺峰,基于支持向量机的近红外光谱技术鉴别掺假牛肉 [J].西北农林科技大学学报(自然科学版),2016,44(12):201-205. |
| [34] |
成甜甜.基于光纤光谱技术的羊肉掺假检测[D].保定:河北农业大学,2021. |
| [35] |
Lopez-Maestresalas A,Insausti K,Jaren C,et al.Detection of minced lamb and beef fraud using NIR spectroscopy [J].Food Control,2019,98:465-473. |
国家自然科学基金项目(3186130281)
甘肃省青年科技基金计划项目(20JR5RA102)
甘肃农业大学伏羲青年英才培养计划项目(GAUfx-02Y07)
/
| 〈 |
|
〉 |