食品科学 ›› 2014, Vol. 35 ›› Issue (6): 156-158.doi: 10.7506/spkx1002-6630-201406033

• 分析检测 • 上一篇    下一篇

基于近红外光谱技术快速识别不同动物源肉品

孟 一1,张玉华1,2,*,王家敏1,2,毕 然1,2,王国利1,2   

  1. 1.山东商业职业技术学院 山东省农产品贮运保鲜技术重点实验室,山东 济南 250103;
    2.国家农产品现代物流工程技术研究中心,山东 济南 250103
  • 收稿日期:2013-06-03 修回日期:2014-02-02 出版日期:2014-03-25 发布日期:2014-04-04
  • 通讯作者: 张玉华 E-mail:z11f@163.com
  • 基金资助:

    济南市高校院所自主创新计划项目(201202060);山东省科技发展计划项目(2012GNC11009)

Rapid Identification of Meat of Different Animal Origins Based on Near Infrared Spectroscopy

MENG Yi1, ZHANG Yu-hua1,2,*, WANG Jia-min1,2, BI Ran1,2, WANG Guo-li1,2   

  1. 1. Shandong Key Laboratory of Storage and Transportation Technology of Agricultural Products, Shandong Institute of Commerce and
    Technology, Jinan 250103, China; 2. National Engineering Research Center for Agricultural Products Logistics,
    Jinan 250103, China
  • Received:2013-06-03 Revised:2014-02-02 Online:2014-03-25 Published:2014-04-04
  • Contact: ZHANG Yu-hua E-mail:z11f@163.com

摘要:

利用傅里叶变换近红外光谱技术建立猪肉、牛肉和羊肉的定性识别模型。用主成分分析法将原始数据压缩为10 个主成分,在全波段9 881.46~4 119.20 cm-1范围内,利用附加散射校正技术结合Savitzky-Golay滤波法对光谱数据进行预处理,采用判别分析法建立模型。模型对训练集的识别准确率为100%,对预测集猪肉、牛肉和羊肉的识别准确率分别为99.28%、97.42%和100%。结果表明:近红外光谱结合模式识别方法能快速无损地识别猪肉、牛肉和羊肉。

关键词: 近红外光谱, 模式识别法, 肉类识别

Abstract:

A qualitative identification model for pork, beef and mutton was established by Fourier transform near-infrared
spectroscopy (NIR). The raw data were compressed to 10 principal components using principal component analysis. The
spectral data were preprocessed by additional scattering correction combined with Savitzky-Golayfilter method in the range
of 9 881.46–4 119.20 cm-1. The model was established using discriminant analysis. Identification accuracy rate of the model
for training set was 100%. Identification accuracy rates of pork, beef and mutton for prediction set were 99.28%, 97.42% and
100%, respectively. These results showed that pork, beef and mutton could be rapidly and non-destructively identified by
near-infrared spectroscopy combined with pattern recognition method.

Key words: near-infrared spectroscopy, pattern recognition method, identification of meat

中图分类号: