Spectroscopy and Spectral Analysis, Volume. 42, Issue 8, 2532(2022)

Study on Geographical Traceability of Artemisia argyi by Employing the Fourier Transform Infrared Spectral Fingerprinting

Chao LI1,*... Meng-zhi LI1,1;, Dan-xia LI1,1;, Shi-bing WEI1,1;, Zhan-hu CUI2,2;, Li-ling XIANG1,1; and Xian-zhang HUANG1,1; *; |Show fewer author(s)
Author Affiliations
  • 11. Henan Key Laboratory of Zhang Zhongjing Formulae and Herbs for Immunoregulation, Nanyang Institute of Technology, Nanyang 473000, China
  • 22. College of Agriculture, Fujian Agriculture and Forestry University, Fuzhou 350002, China
  • show less
    Figures & Tables(12)
    Original spectra of A.argyi from different main producing areas
    Contrast spectrum and characteristic peak of A.argyi
    3D-plots of the first three principal components
    • Table 1. Results of similarity analysis

      View table
      View in Article

      Table 1. Results of similarity analysis

      产地相关系数均值
      共有特征峰选取波段全谱
      河南省南阳市0.9520.9080.931
      河南省安阳市0.9040.8030.734
      湖北省蕲春县111
      浙江省宁波市0.9840.9560.968
      河北省安国市0.9910.9750.988
    • Table 2. Eigenvalues and variance contribution of PCA

      View table
      View in Article

      Table 2. Eigenvalues and variance contribution of PCA

      主成分特征值方差贡献
      值/%
      累计贡献
      率/%
      交叉验证
      率/%
      累计验证
      率/%
      11 442.0982.5082.5082.2082.20
      2167.919.6192.119.3891.58
      357.413.2895.393.0894.66
      447.402.7198.103.1097.76
      512.320.7198.810.6698.42
      69.370.5399.340.6699.08
      73.460.2099.540.2299.30
      83.360.1999.730.2599.55
    • Table 3. Identification effects of KNN algorithm

      View table
      View in Article

      Table 3. Identification effects of KNN algorithm

      预处理方式样品数鉴别正确率/%
      欧氏距离曼哈顿距离夹角余弦
      去噪处理7593.386.786.7
      高斯滤波7593.386.786.7
      归一化处理7593.380.093.3
      多元散射校正7593.380.093.3
      标准正态变换7590.080.080.0
      一阶导数+SG平滑7573.380.073.3
      二阶导数+SG平滑7566.766.760.0
      一阶导数+Norris Gap75100100100
      二阶导数+Norris Gap7566.766.766.7
    • Table 4. Discrimination effects of RF algorithm

      View table
      View in Article

      Table 4. Discrimination effects of RF algorithm

      预处理方式样品数样品归属鉴别正确率/%
      训练集测试集训练集测试集训练集测试集
      去噪处理6015601310086.7
      高斯滤波6015601210080.0
      归一化处理6015601210080.0
      多元散射校正6015601110073.3
      标准正态变换601560810053.3
      一阶导数+SG平滑6015601310086.7
      二阶导数+SG平滑6015601110073.3
      一阶导数+Norris Gap6015601210080.0
      二阶导数+Norris Gap6015601110073.3
    • Table 5. Discrimination effects of Bayes algorithm

      View table
      View in Article

      Table 5. Discrimination effects of Bayes algorithm

      预处理方式样品数样品归属鉴别正确率/%
      训练集测试集训练集测试集训练集测试集
      去噪处理6015601410093.3
      高斯滤波601559998.360.0
      归一化处理6015581296.780.0
      多元散射校正6015581296.780.0
      标准正态变换6015501083.366.7
      一阶导数+SG平滑60156011006.7
      二阶导数+SG平滑60156011006.7
      一阶导数+Norris Gap60156015100100
      二阶导数+Norris Gap60156011006.7
    • Table 6. Identification effects of SVM-pso algorithm

      View table
      View in Article

      Table 6. Identification effects of SVM-pso algorithm

      预处理方式样品数样品归属鉴别正确率/%
      训练集测试集训练集测试集训练集测试集
      去噪处理6015601310086.7
      高斯滤波6015601310086.7
      归一化处理6015601310086.7
      多元散射校正6015601310086.7
      标准正态变换601560710046.7
      一阶导数+SG平滑6015601310086.7
      二阶导数+SG平滑601560810053.3
      一阶导数+Norris Gap60156015100100
      二阶导数+Norris Gap6015601110073.3
    • Table 7. Identification effects of BP neural network algorithm

      View table
      View in Article

      Table 7. Identification effects of BP neural network algorithm

      预处理方式样品数样品归属鉴别正确率/%
      训练集测试集训练集测试集训练集测试集
      去噪处理601560510033.3
      高斯滤波601560810053.3
      归一化处理601560510033.3
      多元散射校正601560710046.7
      标准正态变换601560810053.3
      一阶导数+SG平滑601560910060
      二阶导数+SG平滑601560610040
      一阶导数+Norris Gap60150000
      二阶导数+Norris Gap601560610040
    • Table 8. Discrimination effects of LS-SVM algorithm

      View table
      View in Article

      Table 8. Discrimination effects of LS-SVM algorithm

      预处理方式样品数样品归属鉴别正确率/%
      训练集测试集训练集测试集训练集测试集
      去噪处理601560210013.3
      高斯滤波601560510033.3
      归一化处理601560210013.3
      多元散射校正601560210013.3
      标准正态变换601560210013.3
      一阶导数+SG平滑601560210013.3
      二阶导数+SG平滑601560210013.3
      一阶导数+Norris Gap60156001000
      二阶导数+Norris Gap601560210013.3
    • Table 9. Comparison of different pattern recognition methods

      View table
      View in Article

      Table 9. Comparison of different pattern recognition methods

      模型最佳预处理方式正确率
      /%
      运行时间
      /s
      LS-SVM高斯滤波33.3174.6
      SVM-pso一阶导数+Norris Gap10015.4
      Bayes一阶导数+Norris Gap10063.7
      RF去噪处理86.716.7
      BP-NN一阶导数+SG平滑60.0321.6
      KNN一阶导数+Norris Gap10012.3
    Tools

    Get Citation

    Copy Citation Text

    Chao LI, Meng-zhi LI, Dan-xia LI, Shi-bing WEI, Zhan-hu CUI, Li-ling XIANG, Xian-zhang HUANG. Study on Geographical Traceability of Artemisia argyi by Employing the Fourier Transform Infrared Spectral Fingerprinting[J]. Spectroscopy and Spectral Analysis, 2022, 42(8): 2532

    Download Citation

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

    Category: Orginal Article

    Received: Mar. 23, 2022

    Accepted: --

    Published Online: Mar. 17, 2025

    The Author Email: LI Chao (lichaotcm@126.com)

    DOI:10.3964/j.issn.1000-0593(2022)08-2532-06

    Topics