|本期目录/Table of Contents|

 基于多光谱融合的奶粉掺假诊断方法(PDF)

《纳米技术与精密工程》[ISSN:1672-6030/CN:12-1351/O3]

期数:
2017年5期
页码:
384-388
栏目:
精密测量
出版日期:
2017-09-15

文章信息/Info

Title:
 Adulterated Milk Powder Diagnosis Method Based on Multi-Spectra Fusion
作者:
 陈达 骆文欣 黄志轩 李奇峰
 (天津大学精密仪器与光电子工程学院,天津300072)
Author(s):
 Chen Da Luo Wenxin Huang Zhixuan Li Qifeng
 (School of Precision Instrument and Opto-Electronics Engineering, Tianjin University, Tianjin 300072, China)
关键词:
 拉曼光谱 红外光谱 奶粉掺假 数据融合 DWT-CARS-PLSDA
Keywords:
 Raman spectra infrared spectra milk powder adulteration data fusion DWT-CARS-PLSDA
分类号:
O657.33;TS207.3
DOI:
10.13494/j.npe.20160089
文献标识码:
A
摘要:
 发展了一种多光谱融合新技术,该技术充分利用拉曼光谱与红外光谱的互补特性,并借助数据融合手段,高效实现奶粉掺假检测.为进一步提升数据融合算法的准确性,有机结合离散小波变换(DWT)多尺度特性及竞争性自适应重加权偏最小二乘线性判别(CARS-PLSDA)算法,以有效扣除光谱建模中的干扰信息.为验证多光谱融合技术的有效性,对4种典型奶粉掺假体系分别建立分类判别模型.结果表明,基于DWT-CARS-PLSDA多光谱融合算法所建的面粉、淀粉、糊精和大豆分离蛋白奶粉掺假模型灵敏度分别为94.74%、100%、84.21%和100%,正确率分别为99.42%、98.83%、98.25%和98.83%. 与单独对拉曼光谱或红外光谱建立模型相比,4种模型能够显著提高奶粉掺假检测灵敏度和准确性,为奶粉掺假快速诊断提供了一种有效工具.
Abstract:
 To detect the adulteration of milk power more effectively and more rapidly, a multispectra fusion technique (MFT) consisting of Raman and infrared spectrometries is developed to utilize their complementary molecular characteristics localized in their spectra. In MFT, a data fusion strategy is proposed to combine the discrete wavelet transform (DWT) and competitive adaptive reweighted samplingpartial least squares discriminant analysis (CARS-PLSDA), encoding the multispectra information in the presence of complex spectral interference effectively. The DWT-CARS-PLSDA method is validated and refined by performing MFT analysis of 4 typical milk powder adulterant systems. For each model the satisfactory results are achieved with sensitivity of 94.74%, 100%, 84.21% and 100% respectively and accuracy of 99.42%, 98.83%, 98.25% and 98.83% respectively. The results indicate that the MFT outperformed solitary Raman spectroscopy or infrared spectroscopy in both detection sensitivity and accuracy, revealing that the MFT is a promising tool for discriminating the milk powder adulteration.

参考文献/References

备注/Memo

备注/Memo:
收稿日期: 2016-12-25.
作者简介: 陈 达(1979— ),男,博士,研究员.
通讯作者: 陈达,dachen@tju.edu.cn.
更新日期/Last Update: 2017-09-28