基于拉曼光谱与k最近邻算法的酸奶鉴别

详细信息    查看全文 | 推荐本文 |

英文篇名：Identification of yoghurt based on Raman spectroscopy and k nearest neighbor algorithm 作者：张正勇 ; 岳彤彤 ; 马杰 ; 孙树垒 ; 刘军 ; 王海燕 英文作者：ZHANG Zheng-Yong;YUE Tong-Tong;MA Jie;SUN Shu-lei;LIU Jun;WANG Hai-yan;School of Management Science and Engineering,Nanjing University of Finance and Economics;State Key Laboratory of Chemo/Biosensing and Chemometrics,Hunan University;School of Management Engineering and Electronic Commerce, Zhejiang Gongshang University; 关键词：拉曼光谱 ; 最近邻算法 ; 酸奶 ; 鉴别 英文关键词：Raman spectroscopy;;Nearest neighbor algorithm;;Yogurt;;Identification 中文刊名：FXSY 英文刊名：Chinese Journal of Analysis Laboratory 机构：南京财经大学管理科学与工程学院;湖南大学化学生物传感与计量学国家重点实验室;浙江工商大学管理工程与电子商务学院; 出版日期：2019-04-15 11:30 出版单位：分析试验室 年：2019 期：v.38 基金：国家自然科学基金项目(61602217,71433006,91746202);; 湖南大学化学生物传感与计量学国家重点实验室开放课题基金项目(2017017)资助 语种：中文; 页：FXSY201905009 页数：5 CN：05 ISSN：11-2017/TF 分类号：47-51

摘要

采集了3种品牌酸奶拉曼光谱,单个样品光谱采集时间仅需100 s,通过谱图分析揭示了丰富的分子振动信息,同时发现品牌酸奶谱图相互间具有极高的相似性,传统的人工解谱方法难以实现有效鉴别。进一步建立了基于拉曼光谱的适用于酸奶质量鉴别的优化的k最近邻算法,结果显示在优化条件下,即小波降噪(分解层数N=3,bior2. 4小波基),主成分分析选取前40个主成分(累计贡献率超过95%),k最近邻算法(k=1,马氏距离),本文所建立的快速智能鉴别方法判别时间在1 s之内,即可实现平均识别率99. 70%的鉴别效果。
        The Raman spectra of three brands of yogurt were collected in the experiment,and the spectral collection time of a single sample was only 100 seconds. The rich molecular vibration information was revealed through the spectral analysis. At the same time,it was found that the spectra of different brands yogurt were very similar. The traditional artificial spectral method was difficult to realize the effective identification of them. Therefore,an optimized k nearest neighbor algorithm based on Raman spectroscopy for quality identification of yogurt was established. The results show that under optimal conditions: wavelet denoising( decomposition layer number N = 3,bior2. 4 wavelet base),principal component analysis,the first 40 principal components selected( cumulative contribution rate of more than 95%),k nearest neighbor algorithm( k = 1,Mahalanobis distance),the running time of the established fast-intelligent identification method in this paper was within 1 second,and the average recognition rate was 99. 70%.

引文

[1] Zhang Z Y,Sha M,Wang H Y. J. Raman Spectrosc.,2017,48(8):1111
    [2] Zhang Z Y,Liu J,Wang H Y. Anal. Lett.,2015,48(12):1930
    [3] Hua M Z,Feng S L,Wang S,Lu X N. Food Chem.,2018,258:254
    [4] Rodrigues Júnior P H,de SáOliveira K,Almeida C E R D,De Oliveira L F C,Stephani R,Pinto A F D,Carvalho A F D,PerroneT. Food Chem.,2016,196:584
    [5] Almeida M R,Oliveira K D S,Stephani R,de Oliveira L F C. J. Raman Spectrosc.,2011,42(7):1548
    [6] Liu W H,Yang W,Zhang D. Spectrosc. Spectral Anal.,2008,28(2):343刘文涵,杨未,张丹.光谱学与光谱分析,2008,28(2):343
    [7] Mazurek S,Szostak R,Czaja T,Zachwieja A. Talanta. 2015,138:285
    [8] Mendes T O,Junqueira G M A,Porto B L S,Brito C D,Sato F,de Oliveira M A L,Anjos V,Bell M J V. J. Raman Spectrosc.,2016,47(6):692
    [9] Rygula A,Majzner K,Marzec K M,Kaczor A,Pilarczyk M,Baranska M. J. Raman Spectrosc.,2013,44(8):1061
    [10] Moros J,Garrigues S,de la Guardia M. Anal. Chim. Acta. 2007,593(1):30
    [11] Zhang Z Y,Sha M,Liu J,Wang H Y. China Dairy Ind,2017,45(6):49张正勇,沙敏,刘军,王海燕.中国乳品工业,2017,45(6):49
    [12] Wang H Y,Song C,Sha M,Liu J,Li L P,Zhang Z Y. J. Appl. Spectrosc.,2018,85(2):313
    [13] Wang HY,Gui D D,Sha M,Wang Y B,Chen Y B,Zhang Z Y. China Dairy Cattle,2018(2):55王海燕,桂冬冬,沙敏,王彦波,程永波,张正勇.中国奶牛,2018(2):55
    [14] Wang H Y,Song C,Liu J,Zhang Z Y,Xie W L,Li L P,Sha M. Spectrosc. Spectral Anal.,2017,37(1):124王海燕,宋超,刘军,张正勇,谢伟量,李丽萍,沙敏.光谱学与光谱分析,2017,37(1):124.
    [15] Mei J Y. Research on Mahalanobis distance based metric learning algorithm and its applications[D]. Harbin Institute of Technology,2016梅江元.基于马氏距离的度量学习算法研究及应用[D].哈尔滨工业大学,2016年