• Acta Optica Sinica
  • Vol. 37, Issue 11, 1130003 (2017)
Wensong Wei, Yankun Peng*, Xiaochun Zheng, Wenxiu Wang, and Fang Tian
Author Affiliations
  • National Research & Development Center for Agro-Processing Equipment, College of Engineering, China Agricultural University, Beijing 100083, China
  • show less
    DOI: 10.3788/AOS201737.1130003 Cite this Article Set citation alerts
    Wensong Wei, Yankun Peng, Xiaochun Zheng, Wenxiu Wang, Fang Tian. Rapid Determination of Content of Total Volatile Basic Nitrogen in Pork Based on Multispectral Detection System with Optimal Wavelength[J]. Acta Optica Sinica, 2017, 37(11): 1130003 Copy Citation Text show less
    (a) Hyperspectral image at wavelength of 280 nm of ROI and (b) reflectance curve after calibration of a certain sample
    Fig. 1. (a) Hyperspectral image at wavelength of 280 nm of ROI and (b) reflectance curve after calibration of a certain sample
    Raw reflectance spectra of 52 pork samples
    Fig. 2. Raw reflectance spectra of 52 pork samples
    Spectrograms of raw spectra after preprocessing with FD-SNV algorithm
    Fig. 3. Spectrograms of raw spectra after preprocessing with FD-SNV algorithm
    (a) Number of characteristic wavelength screened by SWA algorithm; (b) wavelength distribution of screening variable
    Fig. 4. (a) Number of characteristic wavelength screened by SWA algorithm; (b) wavelength distribution of screening variable
    Regression coefficient distribution of band screened by SWA algorithm
    Fig. 5. Regression coefficient distribution of band screened by SWA algorithm
    (a) Number of optimal characteristic wavelength screened by SPA algorithm; (b) detailed position of characteristic wavelength
    Fig. 6. (a) Number of optimal characteristic wavelength screened by SPA algorithm; (b) detailed position of characteristic wavelength
    (a) Frequency of characteristic wavelength screened by GA algorithm; (b) modeling contribution rate of characteristic wavelength number
    Fig. 7. (a) Frequency of characteristic wavelength screened by GA algorithm; (b) modeling contribution rate of characteristic wavelength number
    Distribution of characteristic wavelength screened by GA algorithm
    Fig. 8. Distribution of characteristic wavelength screened by GA algorithm
    Distribution of optimal characteristic wavelength screened by three algorithms of SWA, SPA, and GA
    Fig. 9. Distribution of optimal characteristic wavelength screened by three algorithms of SWA, SPA, and GA
    Diagram of multispectral detection system
    Fig. 10. Diagram of multispectral detection system
    Reflectance of 44 pork samples collected by multispectral method
    Fig. 11. Reflectance of 44 pork samples collected by multispectral method
    Predicted content of TVB-N with PLSR model and MLR model. (a) PLSR model, calibration set; (b) PLSR model, prediction set; (c) MLR model, calibration set; (d) MLR model, prediction set
    Fig. 12. Predicted content of TVB-N with PLSR model and MLR model. (a) PLSR model, calibration set; (b) PLSR model, prediction set; (c) MLR model, calibration set; (d) MLR model, prediction set
    SetSample numberMaximum /10-5Minimum /10-5Average /10-5
    Calibration3931.715.5515.45
    Prediction1335.566.5616.85
    Table 1. Measured mass fraction of TVB-N in calibration set and prediction set of the first group pork samples
    SetSample numberMaximum /10-5Minimum /10-5Average /10-5
    Calibration3338.016.9617.69
    Prediction1134.866.2916.09
    Table 2. Measured mass fraction of TVB-N in calibration set and prediction set of the second group pork samples
    Preprocessing methodRcxSEC /10-5RpxSEP /10-5
    Raw spectrum0.87623.44430.86013.9485
    FD0.94662.30260.93722.7969
    SD0.90413.21200.88203.7442
    SNV0.86864.49830.67887.8631
    FD-SNV0.94542.32880.93952.6838
    AS0.80394.36470.76305.1728
    AS-FD0.89554.00740.82515.5752
    AS-SD0.69524.78150.65787.5999
    SD-SNV0.91133.09180.89493.8736
    Table 3. PLSR model results of raw spectra after preprocessing with different methods
    Screening algorithmVariable numberModeling methodRcxSEC /10-5RpxSEP /10-5Screening wavelength /nmt
    SWA5PLSR0.814.180.794.47472,595,765,811,8305/455
    MLR0.834.040.784.62
    SPA13PLSR0.922.880.8923.40452,455,474,532,544,579,586,596,610,635,778,860,86513/455
    MLR0.942.470.8933.14
    GA28PLSR0.952.240.913.78471,481,483,485,489,491,506,525,542,543,544,546,547,548,564,576,593,597,601,605,607,610,612,615,620,807,858,86528/455
    MLR0.962.160.854.24
    Table 4. Modeling results of characteristic wavelength obtained by different screening variable algorithms
    ModelRcxSEC /10-5RpxSEP /10-5
    PLSR0.90144.180.89764.19
    MLR0.90503.630.90403.81
    Table 5. Predicted results of PLSR model and MLR model established by multispectral method
    Wensong Wei, Yankun Peng, Xiaochun Zheng, Wenxiu Wang, Fang Tian. Rapid Determination of Content of Total Volatile Basic Nitrogen in Pork Based on Multispectral Detection System with Optimal Wavelength[J]. Acta Optica Sinica, 2017, 37(11): 1130003
    Download Citation