基于拉曼光谱-模式识别方法对奶粉进行真伪鉴别和掺伪分析

发布时间：2018-03-06 21:30

本文选题：奶粉　切入点：拉曼光谱　出处：《光谱学与光谱分析》2017年01期 　论文类型：期刊论文

【摘要】：奶粉的真伪和掺伪近年来受到广泛的关注,研究一种操作便捷,能准确、快速、全面鉴定奶粉品牌并实现奶粉掺假鉴别的新方法对于奶粉的质量控制具有重要的意义。为实现奶粉的真伪鉴别,采集三种品牌奶粉贝因美、飞鹤和雀巢的拉曼光谱,并利用拉曼谱图特征峰结合最近邻算法(nearest neighbor,NN)的模型对三种品牌奶粉进行识别,在10次交叉验证的基础上,平均识别率为99.56%。为实现奶粉的掺伪分析,将飞鹤奶粉与雀巢奶粉按不同质量比(0∶1,1∶3,1∶1,3∶1,1∶0)混合成五种掺伪奶粉,提取掺伪奶粉中的脂肪,采集脂肪样本的拉曼光谱,分别使用拉曼谱图特征峰结最近邻算法的模型和核主成分分析(kernel principal components analysis,KPCA)结合最近邻算法的模型对五种脂肪样本进行识别,10次交叉验证下的平均识别率分别为93.33%和98.89%,平均运算时间分别为0.085和0.104s。实验证明:特征峰结合NN的算法可以快速实现对奶粉真伪的判别,但此算法不能很好的区分掺伪奶粉;拉曼光谱-KPCA-NN模型可以为奶粉的掺伪检测提供一种简便、准确、快速的方法。
[Abstract]:The authenticity and adulteration of milk powder have received extensive attention in recent years. The new method of comprehensive identification of milk powder brands and identification of milk powder adulteration is of great significance for the quality control of milk powder. In order to realize the identification of milk powder, the Raman spectra of Beyingmei, Feihe and Nestl 茅 of three brands of milk powder were collected. Three brands of milk powder were identified by using Raman spectrum characteristic peak and nearest neighbor algorithm. The average recognition rate was 99.56 on the basis of 10 cross-validations. In order to realize adulteration analysis of milk powder, Flying crane milk powder and Nestle milk powder were mixed into five kinds of adulterated milk powder according to different mass ratio of 0: 1: 1: 1: 1: 1: 1: 1: 1: 1) to extract fat from adulterated milk powder and to collect Raman spectra of fat samples. The average recognition rates of five fat samples were 93.33% and 93.33%, respectively, using the model of Raman spectrum characteristic peak junction nearest neighbor algorithm and kernel principal component analysis kernel principal components analysis KPCA combined with nearest neighbor algorithm. The average operation time is 0.085 and 0.104s respectively. The experiment shows that the algorithm of characteristic peak combined with NN can quickly distinguish the authenticity of milk powder. However, this algorithm can not distinguish adulterated milk powder well, and the Raman spectrum -KPCA-NN model can provide a simple, accurate and fast method for the detection of adulteration of milk powder.
【作者单位】：南京财经大学管理科学与工程学院;江苏省质量安全工程研究院;南京理工大学机械工程学院;
【基金】：质检公益性行业科研专项(201410173) 国家重大科学仪器设备开发专项(2013YQ090703) 国家自然科学基金项目(61373058)资助
【分类号】：TS252.7;O657.37

【相似文献】