Laser & Optoelectronics Progress, Volume. 55, Issue 5, 053004(2018)

Spectral Reflectance Reconstruction Based on Dimension Reduction Regularization Polynomials

Ke Wang1,2、1; 2; , Huiqin Wang1,2、1; 2; , Yanqun Long1、2; , Weichao Wang1、2; , Lijuan Zhao1、2; , and Lei Yang1、2;
Author Affiliations
  • 1 School of Information and Control Engineering, Xi'an University of Architecture and Technology, Xi'an, Shaanxi 710055, China
  • 1 School of Management, Xi'an University of Architecture and Technology, Xi'an, Shaanxi 710055, China
  • 2 School of Information and Control Engineering, Xi'an University of Architecture and Technology, Xi'an, Shaanxi 710055, China
  • show less
    Figures & Tables(13)
    Multispectral camera
    Transmissivity curve of filter
    Relationship between training sample number and mean RMSE of reconstructed spectral reflectance
    Regularization parameters obtained by L-curve method. (a) Training result 1; (b) training result 2
    RMSE of 20 testing samples with three reconstruction methods
    Spectral reflectance curves with three reconstruction methods. (a) xISSD=0.022; (b) xISSD=0.121; (c) xISSD=0.351
    Mural referential color patches and multispectral images. (a) Markings of mural referential color patches; (b) multispectral images with 11 channels
    CIELAB chromaticity distribution space of mural referential color patches obtained by different reconstruction methods
    Reconstructed and measured spectral reflectance curves of six referential color patches of mural. (a) Color patch 1; (b) color patch 2; (c) color patch 3; (d) color patch 4; (e) color patch 5; (f) color patch 6
    • Table 1. CVC, color difference, and RMSE with different principal component numbers

      View table

      Table 1. CVC, color difference, and RMSE with different principal component numbers

      kCVC /%Mean ΔEMean RMSE
      598.870.7170.082
      699.950.4300.058
      799.890.4240.042
      899.990.2630.033
    • Table 2. Color difference and RMSE with different polynomial term numbers

      View table

      Table 2. Color difference and RMSE with different polynomial term numbers

      nMean ΔEMean RMSE
      60.4820.032
      90.3830.023
      120.3610.020
      140.3450.020
      180.3410.019
    • Table 3. Spectral reflectance reconstruction accuracies of PCA, PRE and DRRP methods

      View table

      Table 3. Spectral reflectance reconstruction accuracies of PCA, PRE and DRRP methods

      MethodRMSEGFC /%ISSDΔE
      MeanMinimumMaximumMeanMinimumMaximumMeanMinimumMaximumMeanMinimumMaximum
      PCA0.0580.0120.10698.2093.5399.630.1460.0640.6370.4300.1720.931
      PRE0.0200.0110.04298.1792.5399.740.1600.0210.5560.3610.0640.766
      DRRP0.0180.0070.03299.5198.8199.960.1060.0110.3590.2830.0600.412
    • Table 4. Spectral reflectance reconstruction accuracies of six referential color patches of mural

      View table

      Table 4. Spectral reflectance reconstruction accuracies of six referential color patches of mural

      Number of referentialcolor patchesDRRP methodPRE methodPCA method
      RMSEGFC /%ISSDRMSEGFC /%ISSDRMSEGFC /%ISSD
      10.012399.630.09310.014598.430.14290.056297.780.1325
      20.011299.210.15680.017898.150.20470.059397.010.1930
      30.031799.180.16130.034798.070.21630.071296.810.2013
      40.019499.420.12360.023998.210.17450.067497.230.1701
      50.037699.090.17020.046298.820.22130.082097.860.2113
      60.011599.780.10720.017798.650.15860.048998.010.1489
      Mean0.020699.390.13530.025898.390.18640.067597.450.1761
    Tools

    Get Citation

    Copy Citation Text

    Ke Wang, Huiqin Wang, Yanqun Long, Weichao Wang, Lijuan Zhao, Lei Yang. Spectral Reflectance Reconstruction Based on Dimension Reduction Regularization Polynomials[J]. Laser & Optoelectronics Progress, 2018, 55(5): 053004

    Download Citation

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

    Category: Spectroscopy

    Received: Oct. 20, 2017

    Accepted: --

    Published Online: Sep. 11, 2018

    The Author Email: Wang Ke (wangke@xauat.edu.cn)

    DOI:10.3788/LOP55.053004

    Topics