• Spectroscopy and Spectral Analysis
  • Vol. 41, Issue 2, 565 (2021)
Ya-xuan WANG1、1、*, Feng TAN1、1, Yuan-ming XIN1、1, Huan LI1、1, Xiao-yu ZHAO1、1, and Bao-xin LU1、1
Author Affiliations
  • 1[in Chinese]
  • 11. College of Civil Engineering and Water Conservancy, Helongjiang Bayi Agricultural University, Daqing 163319, China
  • show less
    DOI: 10.3964/j.issn.1000-0593(2021)02-0565-07 Cite this Article
    Ya-xuan WANG, Feng TAN, Yuan-ming XIN, Huan LI, Xiao-yu ZHAO, Bao-xin LU. Identification of Rice From Similar Areas With Different Pretreatment Methods of Raman Spectrum[J]. Spectroscopy and Spectral Analysis, 2021, 41(2): 565 Copy Citation Text show less
    Raw Raman spectrum of three producing area of rice
    Fig. 1. Raw Raman spectrum of three producing area of rice
    Main characteristic peaks of rice Raman spectrum
    Fig. 2. Main characteristic peaks of rice Raman spectrum
    Pre-processing method of first derivative+translation smoothing
    Fig. 3. Pre-processing method of first derivative+translation smoothing
    Pre-processing method of second derivative+translation smoothing
    Fig. 4. Pre-processing method of second derivative+translation smoothing
    Pre-processing method of wavelet transform+baseline removal
    Fig. 5. Pre-processing method of wavelet transform+baseline removal
    Pre-processing method of piecewise polynomial fitting+baseline removal
    Fig. 6. Pre-processing method of piecewise polynomial fitting+baseline removal
    Relationship between true value and predicted value
    Fig. 7. Relationship between true value and predicted value
    序号预处理方法训练集测试集
    CCMSERMSECCMSERMSE
    1一阶导数+平移平滑0.993 80.002 70.052 20.935 70.027 70.166 4
    2二阶导数+平移平滑0.943 40.024 40.156 30.839 80.065 50.255 9
    3小波变换+去除基线0.885 60.047 90. 218 60.884 60.048 40.220 0
    43点2次拟合+去除基线0.990 00.004 40.066 50.960 60.017 20.131 1
    54点2次拟合+去除基线0.994 60.002 40.048 90.887 00.050 50.224 7
    65点2次拟合+去除基线0.954 80.019 60.140 10.690 60.154 50.393 0
    73点3次拟合+去除基线0.981 70.008 10.089 90.929 70.030 20.173 7
    83点4次拟合+去除基线0.977 60.009 80.099 20.923 90.032 60.180 5
    93点5次拟合+去除基线0.976 30.010 40.102 00.896 00.044 20.210 3
    Table 1. Correlation coefficient, Mean square error, Root mean square error of different pretreatment methods
    序号预处理方法ABC总识别率/%
    识别数识别率/%识别数识别率/%识别数识别率/%
    1一阶导数+平移平滑171001482.41482.488.2
    2二阶导数+墙移平滑1694.11482.41482.486.2
    3小波变换+平移平滑171001588.21710096.1
    43点2次拟合+去除基线171001710017100100
    54点2次拟合+去除基线171001482.41588.290.2
    65点2次拟合+去除基线1588.2741.1529.452.9
    73点3次拟合+去除基线1694.11694.11710096.1
    83点4次拟合+去除基线1694.1171001710098.0
    93点5次拟合+去除基线1694.1171001694.196.1
    Table 2. The number and recognition rate of different pre-processing methods in 17 test sample
    Ya-xuan WANG, Feng TAN, Yuan-ming XIN, Huan LI, Xiao-yu ZHAO, Bao-xin LU. Identification of Rice From Similar Areas With Different Pretreatment Methods of Raman Spectrum[J]. Spectroscopy and Spectral Analysis, 2021, 41(2): 565
    Download Citation