• Spectroscopy and Spectral Analysis
  • Vol. 41, Issue 12, 3665 (2021)
Yan-ling CHEN*, Liang-lun CHENG*;, Heng WU*;, Li-min XU, Wei-jian HE, and Feng LI
Author Affiliations
  • School of Computer, Guangdong University of Technology, Guangzhou 510006, China
  • show less
    DOI: 10.3964/j.issn.1000-0593(2021)12-3665-06 Cite this Article
    Yan-ling CHEN, Liang-lun CHENG, Heng WU, Li-min XU, Wei-jian HE, Feng LI. A Method of Terahertz Spectrum Material Identification Based on Wavelet Coefficient Graph[J]. Spectroscopy and Spectral Analysis, 2021, 41(12): 3665 Copy Citation Text show less
    TAS7400 terahertz time domain spectrometer
    Fig. 1. TAS7400 terahertz time domain spectrometer
    Convolutional neural network diagram
    Fig. 2. Convolutional neural network diagram
    Terahertz spectral curves and wavelet transform graph
    Fig. 3. Terahertz spectral curves and wavelet transform graph
    类别测试样品
    含吸收峰乳糖、 葡萄糖、 果糖、 蔗糖、 麦芽糖、 山梨酸钾、 苯甲酸、 苯甲酸钠、 叔丁基氢醌、 滑石粉
    无明显吸收峰红磷、 淀粉、 乙基麦芽酚、 甲基环戊烯醇酮、 溴化钾、 氯化钾、 氯化钙、 硫酸钾、 硫酸钠、 聚乙烯
    Table 1. Summary of experimental sample
    第一组第二组第三组
    CNNa100%98.13%97.19%
    b100%100%100%
    SVMa32.22%16.67%20.27%
    b76.39%58.33%62.50%
    MLPa94.44%90.27%81.94%
    b97.22%94.44%86.11%
    Table 2. Summary of pure substances recognition results
    小波寻峰识别本文
    识别率平均相似度识别率平均相似度
    麦芽糖90%61.17%100%97.14%
    果糖35%38.51%92.86%89.51%
    乳糖100%95.07%92.86%97.33%
    葡萄糖100%71.69%100%97.60%
    蔗糖25%38.88%100%91.69%
    淀粉--100%89.44%
    Table 3. The recognition rate summary of the two methods
    Yan-ling CHEN, Liang-lun CHENG, Heng WU, Li-min XU, Wei-jian HE, Feng LI. A Method of Terahertz Spectrum Material Identification Based on Wavelet Coefficient Graph[J]. Spectroscopy and Spectral Analysis, 2021, 41(12): 3665
    Download Citation