• Spectroscopy and Spectral Analysis
  • Vol. 42, Issue 1, 282 (2022)
Xiao-kang ZHAO*, Xin ZHAO, Qi-bing ZHU*;, and Min HUANG
Author Affiliations
  • Key Laboratory of Advanced Process Control for Light Industry (Ministry of Education), Jiangnan University, Wuxi 214122, China
  • show less
    DOI: 10.3964/j.issn.1000-0593(2022)01-0282-10 Cite this Article
    Xiao-kang ZHAO, Xin ZHAO, Qi-bing ZHU, Min HUANG. A Model Construction Method of Spectral Nondestructive Detection for Apple Quality Based on Unsupervised Active Learning[J]. Spectroscopy and Spectral Analysis, 2022, 42(1): 282 Copy Citation Text show less
    Flow chart of spectral detecting method based on HAC-LLR training samples selecting strategy
    Fig. 1. Flow chart of spectral detecting method based on HAC-LLR training samples selecting strategy
    The average spectra of three cultivars apple samples harvestee from two years
    Fig. 2. The average spectra of three cultivars apple samples harvestee from two years
    PLSR mdoel prediction results of SSC (a) and firmness (b) based on different sample selection algorithms under different datsets
    Fig. 3. PLSR mdoel prediction results of SSC (a) and firmness (b) based on different sample selection algorithms under different datsets
    Normalized AUCs of the RMSE (a), the Rp (b) and the RPD (c) on different datasets
    Fig. 4. Normalized AUCs of the RMSE (a), the Rp (b) and the RPD (c) on different datasets
    收获年份品种数量SSC/%硬度/N
    均值方差最小值最大值均值方差最小值最大值
    GD1 07014.8141.4849.70018.80055.54914.80030.43198.893
    2009JG87412.7191.0599.40016.40062.95122.17329.319110.187
    RD1 07811.5581.1608.10015.40058.91115.09929.49789.980
    GD113113.2321.2709.60017.50063.28016.69829.434100.994
    2010JG1 08712.5451.5328.70017.30063.40519.01527.700105.340
    RD1 03512.2521.4548.80015.50069.06816.65429.978107.428
    Table 1. Statistics of quality reference for apple samples
    品种年份算法预测集
    SSC硬度
    RMSERpRPDDif./%RMSERpRPDDif./%
    GD2009RS0.8240.8732.005.99.0570.8251.7653.3
    KS0.8110.8722.0374.48.8830.8281.8121.4
    SPXY0.8420.8731.9718.09.3990.8221.7196.8
    HAC-LLR0.7750.8822.1228.7600.8351.831
    JG2009RS0.6830.7991.6492.08.6010.9242.5885.1
    KS0.7260.7641.5567.98.3530.9282.6722.3
    SPXY0.7310.7851.5508.58.7460.9212.5526.7
    HAC-LLR0.6690.8081.6928.1640.9312.732
    RD2009RS0.8020.7761.5058.59.2570.8141.7024.3
    KS0.7680.7761.5724.49.0430.8181.7452.1
    SPXY0.7950.7811.5137.79.0970.8171.7322.6
    HAC-LLR0.7340.8001.6508.8560.8271.778
    Table 2. The prediction results of PLSR models based on 200 samples from 2009 selected by four algorithms respectively
    品种年份算法预测集
    SSC硬度
    RMSERpRPDDif./%RMSERpRPDDif./%
    GD2010RS0.6490.8812.1108.68.2140.8742.0502.6
    KS0.6320.8872.1776.28.0960.8822.0761.2
    SPXY0.6420.8862.1337.68.3490.8742.0134.2
    HAC-LLR0.5930.9002.3158.0000.8822.101
    JG2010RS0.7590.8872.1655.011.0830.8281.7644.2
    KS0.7580.8902.1664.911.4330.8171.7077.2
    SPXY0.7420.9012.2112.811.1940.8291.7435.2
    HAC-LLR0.7210.8992.27910.6150.8371.845
    RD2010RS0.7480.8722.0118.09.5240.8491.8507.2
    KS0.7140.8802.1043.69.2290.8501.9174.2
    SPXY0.7930.8581.90213.210.4860.8241.69115.7
    HAC-LLR0.6880.8902.1848.8370.8651.991
    Table 3. The prediction results of PLSR models based on 200 samples from 2010 selected by four algorithms respectively
    Xiao-kang ZHAO, Xin ZHAO, Qi-bing ZHU, Min HUANG. A Model Construction Method of Spectral Nondestructive Detection for Apple Quality Based on Unsupervised Active Learning[J]. Spectroscopy and Spectral Analysis, 2022, 42(1): 282
    Download Citation