• Laser & Optoelectronics Progress
  • Vol. 57, Issue 1, 013001 (2020)
Ya He and Jifen Wang*
Author Affiliations
  • Institute of Forensic Science and Technology, People's Public Security University of China, Beijing, 100038, China
  • show less
    DOI: 10.3788/LOP57.013001 Cite this Article Set citation alerts
    Ya He, Jifen Wang. Rapid Nondestructive Identification of Wood Lacquer Using Raman Spectroscopy Based on Characteristic-Band-Fisher-K Nearest Neighbor[J]. Laser & Optoelectronics Progress, 2020, 57(1): 013001 Copy Citation Text show less
    Preprocessing results of Raman spectra of CY003 samples. (a) BL; (b) nine-point smoothing; (c) first derivative; (d) second derivative
    Fig. 1. Preprocessing results of Raman spectra of CY003 samples. (a) BL; (b) nine-point smoothing; (c) first derivative; (d) second derivative
    Classification of three brands of woodlacquer samples obtained by characteristic-band ratio method
    Fig. 2. Classification of three brands of woodlacquer samples obtained by characteristic-band ratio method
    Discriminant distribution of three brands of woodlacquer samples obtained by functions 1 and 2
    Fig. 3. Discriminant distribution of three brands of woodlacquer samples obtained by functions 1 and 2
    NumberBrandCategoryManufacturer
    CY003CYSOWater-basedmultifunctional paintCYSO
    CY008CYSOWater-basedmultifunctional paintCYSO
    HC004HUACAIWater-basedpaintChaomeiyaqi
    HC013HUACAIWater-basedpaintChaomeiyaqi
    QS001TRYKONWaterbornefurniture finishChangfeng
    QS010TRYKONWaterbornefurniture finishChangfeng
    Table 1. Basic information of six samples of wood lacquer
    IndexOriginal spectrumBL+SmoothingBL+Smoothing+1st derivativeBL+Smoothing+2nd derivative
    Overall accuracy /%68.079.896.0100
    Number of errors553570
    Cross validation rate /%65.169.494.1100
    Table 2. Discriminant results of different preprocessing methods
    FunctionEigen valueCumulative varianceRegular correlationTest functionWilks’ LambdaSignificance
    1113.87190.10.9961 and 20.0010
    212.4551000.96210.0740.041
    Table 3. Summary of Fisher-function discriminant
    Spectral typeOverall accuracy /%Cross validationrate /%
    FisherKNN
    BL+Smoothing+1st derivative9679.2100
    BL+Smoothing+2nd derivative10088.5100
    Table 4. Classification results of samples for two models
    Ya He, Jifen Wang. Rapid Nondestructive Identification of Wood Lacquer Using Raman Spectroscopy Based on Characteristic-Band-Fisher-K Nearest Neighbor[J]. Laser & Optoelectronics Progress, 2020, 57(1): 013001
    Download Citation