• Spectroscopy and Spectral Analysis
  • Vol. 42, Issue 6, 1792 (2022)
Yan-de LIU* and Shun WANG
Author Affiliations
  • School of Mechatronics & Vehicle Engineering, East China Jiaotong University, Nanchang 330013, China
  • show less
    DOI: 10.3964/j.issn.1000-0593(2022)06-1792-06 Cite this Article
    Yan-de LIU, Shun WANG. Research on Non-Destructive Testing of Navel Orange Shelf Life Imaging Based on Hyperspectral Image and Spectrum Fusion[J]. Spectroscopy and Spectral Analysis, 2022, 42(6): 1792 Copy Citation Text show less
    Hyperspectral image of navel orange sample(a): Day 0 navel orange; (b): Day 7 navel orange; (c): Day 14 navel orange
    Fig. 1. Hyperspectral image of navel orange sample
    (a): Day 0 navel orange; (b): Day 7 navel orange; (c): Day 14 navel orange
    Comparison chart of representative spectra of navel oranges with different shelf life
    Fig. 2. Comparison chart of representative spectra of navel oranges with different shelf life
    The prediction set classification results of RBF-kernel in LS-SVM
    Fig. 3. The prediction set classification results of RBF-kernel in LS-SVM
    The prediction set classification results of LIN-kernel in LS-SVM
    Fig. 4. The prediction set classification results of LIN-kernel in LS-SVM
    The prediction set classification results of LIN-kernel in LS-SVM
    Fig. 5. The prediction set classification results of LIN-kernel in LS-SVM
    输入变
    量个数
    RMSEPRp误判率
    /%
    预测集误判个数
    第0天第7天第14天
    1760.290.948303
    Table 1. PLS-DA model results based on spectral characteristics
    输入变
    量个数
    核函数参数总误判
    率/%
    预测集误判个数
    第0天第7天第14天
    176LIN-Kernelγ=1.49.3151
    176RBF-Kernelγ=59 078,
    σ2=6 359
    5.33211
    Table 2. LS-SVM model results based on spectral characteristics
    输入变
    量个数
    RMSEPRp误判率
    /%
    预测集误判个数
    第0天第7天第14天
    110.2380.8821.3196
    Table 3. Results of PLS-DA model based on image features
    输入变
    量个数
    核函数参数总误判
    率/%
    预测集误判个数
    第0天第7天第14天
    11LIN-Kernelγ=3 79520096
    11RBF-Kernelγ=124,
    σ2=177
    22.20116
    Table 4. LS-SVM model results based on image features
    输入变
    量个数
    RMSEPRp误判率
    /%
    预测集误判个数
    第0天第7天第14天
    1870.20.971.3010
    Table 5. Results of PLS-DA model based on mixed features
    输入变
    量个数
    核函数参数总误判
    率/%
    预测集误判个数
    第0天第7天第14天
    187LIN-Kernelγ=8.51.33010
    187RBF-Kernelγ=24 810,
    σ2=7 595
    2.67020
    Table 6. LS-SVM model results based on mixed features
    模型不同类
    型特征
    变量
    个数
    核函数参数预测集误
    判率/%
    LS-SVM光谱176RBF-Kernelγ=59 078,
    σ2=6 359
    5.3
    LS-SVM图像11LIN-Kernelγ=3 79520
    LS-SVM融合187LIN-Kernelγ=8.51.33
    Table 7. Result statistics of two qualitative discriminant models with different characteristics
    Yan-de LIU, Shun WANG. Research on Non-Destructive Testing of Navel Orange Shelf Life Imaging Based on Hyperspectral Image and Spectrum Fusion[J]. Spectroscopy and Spectral Analysis, 2022, 42(6): 1792
    Download Citation