Spectroscopy and Spectral Analysis, Volume. 41, Issue 8, 2437(2021)

Study on Characteristic Wavelength Extraction Method for Near Infrared Spectroscopy Identification Based on Genetic Algorithm

Hao-guang LI*, Yun-hua YU1; 2;, Yan PANG1;, and Xue-feng SHEN1; 2;
Author Affiliations
  • 1. College of Mechanical and Control Engineering, Shandong Institute of Petrochemical and Chemical Technology, Dongying 257061, China
  • show less
    Figures & Tables(13)
    Selection of wavelength points based on genetic algorithms
    Recognition rate as population size increasing
    Curve of convergence time as population size increasing
    Curve of recognition rate as crossing rate changing
    Curve of convergence time as crossover rate changing
    Curve of recognition rate as variation rate changing
    Curve of convergence time as variation rate changing
    Curve of recognition rate when characteristic wavelength points increasing
    • Table 1. Characteristic wavelength point list (nm)

      View table
      View in Article

      Table 1. Characteristic wavelength point list (nm)

      12345678910
      9701 1001 1991 217.81 279.81 304.51 329.31 335.51 515.11 639
    • Table 2. Recognition results (Characteristic wavelength point)

      View table
      View in Article

      Table 2. Recognition results (Characteristic wavelength point)

      测试集T1T2T3T4T5
      正确识别率/%94.894.495.691.291.8
      正确拒识率/%94.693.298.093.293.2
      平均识别率/%94.793.896.892.292.5
    • Table 3. Recognition results (whole spectra)

      View table
      View in Article

      Table 3. Recognition results (whole spectra)

      测试集T1T2T3T4T5
      正确识别率/%95.095.095.090.0100.0
      正确拒识率/%95.095.0100.095.090.0
      平均识别率/%95.095.097.592.595.0
    • Table 4. Recognition results (Characteristic wavelength point)

      View table
      View in Article

      Table 4. Recognition results (Characteristic wavelength point)

      测试集T2T3T4T5
      正确识别率/%96.094.096.093.0
      正确拒识率/%96.094.091.092.0
      平均识别率/%96.094.094.092.5
    • Table 5. Recognition results (whole spectra)

      View table
      View in Article

      Table 5. Recognition results (whole spectra)

      测试集T2T3T4T5
      正确识别率/%95.095.096.093.0
      正确拒识率/%98.094.092.094.0
      平均识别率/%96.594.594.093.5
    Tools

    Get Citation

    Copy Citation Text

    Hao-guang LI, Yun-hua YU, Yan PANG, Xue-feng SHEN. Study on Characteristic Wavelength Extraction Method for Near Infrared Spectroscopy Identification Based on Genetic Algorithm[J]. Spectroscopy and Spectral Analysis, 2021, 41(8): 2437

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category: Research Articles

    Received: Jan. 19, 2021

    Accepted: --

    Published Online: Sep. 8, 2021

    The Author Email: LI Hao-guang (lihaoguang@upc.edu.cn)

    DOI:10.3964/j.issn.1000-0593(2021)08-2437-06

    Topics