• Laser & Optoelectronics Progress
  • Vol. 60, Issue 1, 0130004 (2023)
Dongyan Zhang, Congcong Fu*, Dandan Li, Miaoyuan Ma, and Ying Huang
Author Affiliations
  • College of Mechanical and Electrical Engineering, Northeast Forestry University, Harbin 150040, Heilongjiang, China
  • show less
    DOI: 10.3788/LOP213222 Cite this Article Set citation alerts
    Dongyan Zhang, Congcong Fu, Dandan Li, Miaoyuan Ma, Ying Huang. Nondestructive Detection Model of Hazelnut Protein Based on Near Infrared Spectroscopy[J]. Laser & Optoelectronics Progress, 2023, 60(1): 0130004 Copy Citation Text show less
    Schematic diagram of the hardware platform
    Fig. 1. Schematic diagram of the hardware platform
    Experimental flow chart
    Fig. 2. Experimental flow chart
    Raw spectra of all bands
    Fig. 3. Raw spectra of all bands
    Spectra after preprocessing
    Fig. 4. Spectra after preprocessing
    Optimal modeling result of the CARS algorithm. (a) Number of wavebands; (b) RMSEc
    Fig. 5. Optimal modeling result of the CARS algorithm. (a) Number of wavebands; (b) RMSEc
    Variation curve of RMSE of SPA with the number of wavebands
    Fig. 6. Variation curve of RMSE of SPA with the number of wavebands
    Location of selected bands for SPA
    Fig. 7. Location of selected bands for SPA
    Selection probability of each band of the IRF algorithm
    Fig. 8. Selection probability of each band of the IRF algorithm
    Variation curve of RMSEc of the IRF algorithm with the number of characteristic bands
    Fig. 9. Variation curve of RMSEc of the IRF algorithm with the number of characteristic bands
    Regression results of the IRF algorithm on the cross-validation set
    Fig. 10. Regression results of the IRF algorithm on the cross-validation set
    Regression results of the IRF algorithm on the test set
    Fig. 11. Regression results of the IRF algorithm on the test set
    Data setNumberRange /%Mean /%Standard deviation /%
    Hazelnut sample set30020.10-26.9323.501.58
    Table 1. Content distribution of protein in hazelnut
    MethodCross validation setTest set
    RMSEcRcRMSEpRp
    PLS model before pretreatment0.07920.51830.04840.6089
    PLS model after pretreatment0.05500.75900.03650.6374
    Table 2. Prediction results of the model before and after preprocessing
    AlgorithmCross validation setTest set
    RMSEcRcRMSEpRp
    Pretreatment0.05500.75900.03650.6374
    CARS0.03960.89720.01810.8840
    SPA0.03150.89530.03700.7963
    RF0.01860.89220.04030.8876
    IRF0.01780.90820.03720.8999
    Table 3. Modeling results of different algorithms
    Dongyan Zhang, Congcong Fu, Dandan Li, Miaoyuan Ma, Ying Huang. Nondestructive Detection Model of Hazelnut Protein Based on Near Infrared Spectroscopy[J]. Laser & Optoelectronics Progress, 2023, 60(1): 0130004
    Download Citation