食品科学 ›› 2018, Vol. 39 ›› Issue (12): 295-300.doi: 10.7506/spkx1002-6630-201812045

• 安全检测 • 上一篇    下一篇

基于线扫描拉曼高光谱系统的奶粉中硫氰酸钠无损检测

刘宸1,2,3,4,杨桂燕2,3,4,王庆艳2,3,4,黄文倩2,3,4,王超鹏1,2,3,4,王晓彬2,3,4,陈立平1,2,3,4,*   

  1. (1.西北农林科技大学机械与电子工程学院,陕西?杨凌 712100;2.国家农业智能装备工程技术研究中心,北京 100097;3.农业部农业信息技术重点实验室,北京 100097;4.农业智能装备技术北京市重点实验室,北京 100097)
  • 出版日期:2018-06-25 发布日期:2018-06-15
  • 基金资助:
    国家自然科学基金青年科学基金项目(61605009);北京市科技新星计划项目(Z161100004916076); 北京市优秀人才项目(2015000021223ZK40)

Non-Invasive Detection of Sodium Thiocyanate in Milk Powder Using Line-Scanning Raman Hyperspectral Imaging System

LIU Chen1,2,3,4, YANG Guiyan2,3,4, WANG Qingyan2,3,4, HUANG Wenqian2,3,4, WANG Chaopeng1,2,3,4, WANG Xiaobin2,3,4, CHEN Liping1,2,3,4,*   

  1. (1. College of Mechanical and Electronic Engineering, Northwest A&F University, Yangling 712100, China;2. National Research Center of Intelligent Equipment for Agriculture, Beijing 100097, China;3. Key Laboratory of Agri-Informatics, Ministry of Agriculture, Beijing 100097, China;4. Beijing Key Laboratory of Intelligent Equipment Technology for Agriculture, Beijing 100097, China)
  • Online:2018-06-25 Published:2018-06-15

摘要: 搭建一套线扫描拉曼高光谱成像系统,探索针对大面积奶粉样本的硫氰酸钠快速无损检测。首先测量硫氰酸钠产生的拉曼信号在脱脂奶粉层中的穿透情况,然后制备10?种不同硫氰酸钠质量分数的奶粉混合样本并采集拉曼高光谱图像,通过高斯窗平滑法和自适应迭代惩罚最小二乘基线校正方法对拉曼光谱进行预处理。预处理后提取2?068.48?cm-1位移处的单波段图像进行分析,结合二值图像最终获得了样本中硫氰酸钠颗粒的含量以及空间分布。结果显示,在2?068.48?cm-1单波段图像中,感兴趣区域内所有像素点的拉曼强度平均值随着硫氰酸钠含量的增加呈线性增长,其决定系数R2达到了0.991?5。二值图像中,感兴趣区域内所有硫氰酸钠检测点之和呈指数增长趋势。在本实验方法中,单次检测奶粉样本的总面积达到80?mm×80?mm,检测时不接触、不破坏样本,无需借助化学试剂。奶粉混合样本中硫氰酸钠颗粒的检测限可达0.01%。研究结果表明,拉曼高光谱成像系统能够快速、无损且大面积地检测出奶粉中的硫氰酸钠,并且可以直观地展示硫氰酸钠颗粒的具体分布,在实际检测应用中该方法具有巨大潜力。

关键词: 高光谱成像技术, 拉曼光谱, 脱脂奶粉, 硫氰酸钠, 线扫描

Abstract: In this study, a line-scanning Raman hyperspectral imaging system was built to quantify sodium thiocyanate illegally added in skim milk powders with large sample areas. Firstly, the penetration depth of Raman signal produced by sodium thiocyanate in skim milk powders was measured. Then, ten milk powder mixtures with different concentrations of sodium thiocyanate were prepared and their hyperspectral images were also collected. The Gaussian filter smoothing and the airPLS baseline correction methods were used to preprocess the Raman spectra. The corrected images at 2 068.48 cm-1 waveband were extracted to determine the presence of sodium thiocyanate particles. The concentration and spatial distribution of the sodium thiocyanate particles were finally obtained using a simple binarization method. The results showed that the average Raman intensity of all the pixels in the region of interest increased linearly with the increase of sodium thiocyanate particles, and the determination coefficient was 0.991 5. In the binary images, the sum of all sodium thiocyanate pixels in the region of interest increased exponentially. The limit of detection (LOD) for sodium thiocyanate concentration was estimated as 0.01% without using any chemical reagents. This system could expand the area of milk powder samples to 80 mm × 80 mm at a time in a non-invasive and non-contact way. In our research, the sodium thiocyanate particles could be fast, non-destructively and extensively detected using the line-scanning Raman hyperspectral imaging system. Both the concentration and the spatial distribution of the sodium thiocyanate particles in skim milk powder could be acquired at the same time. This method has a great potential for practical application.

Key words: hyperspectral imaging, Raman spectroscopy, skim milk powder, sodium thiocyanate, line-scanning

中图分类号: