• Spectroscopy and Spectral Analysis
  • Vol. 41, Issue 11, 3405 (2021)
Kai QIN1、1;, Gang CHEN2、2;, Jian-yi ZHANG1、1; 2;, and Xia-ping FU1、1; *;
Author Affiliations
  • 11. Faculty of Mechanical Engineering and Automation, Zhejiang Sci-Tech University, Hangzhou 310018, China
  • 22. Zhejiang DEKFELLER Intelligent Machinery Manufacturing Co., Ltd., Hangzhou 310014, China
  • show less
    DOI: 10.3964/j.issn.1000-0593(2021)11-3405-06 Cite this Article
    Kai QIN, Gang CHEN, Jian-yi ZHANG, Xia-ping FU. Optimization of Fruit Pose and Modeling Method for Online Spectral Detection of Apple Moldy Core[J]. Spectroscopy and Spectral Analysis, 2021, 41(11): 3405 Copy Citation Text show less
    Contrast of healthy fruit (left) and fruit with moldy core (right)
    Fig. 1. Contrast of healthy fruit (left) and fruit with moldy core (right)
    Average transmission spectra of healthy fruit and fruit with moldy core in four transportation postures(a): Stem upward; (b): Stem downward; (c): Stem towards the transportation direction;(d): Stem perpendicular to the transportation direction
    Fig. 2. Average transmission spectra of healthy fruit and fruit with moldy core in four transportation postures
    (a): Stem upward; (b): Stem downward; (c): Stem towards the transportation direction;(d): Stem perpendicular to the transportation direction
    Comparison of the performance (a) and the number of moldy core apples incorrectly identified as healthy (b) of the six discriminant models
    Fig. 3. Comparison of the performance (a) and the number of moldy core apples incorrectly identified as healthy (b) of the six discriminant models
    VIP score distribution of PLS-DA model
    Fig. 4. VIP score distribution of PLS-DA model
    样品分类数量
    /个
    平均最大
    果径/mm
    平均高度
    /mm
    平均重量
    /g
    健康果7887.071.5240.3
    病果1886.470.7224.6
    全部样品9686.971.4237.4
    Table 1. Appearance parameter statistics of healthy fruit and fruit with moldy core
    方法摆放姿态PCs训练集预测集
    错判个数准确率
    /%
    错判个数准确率
    /%
    病果健康果病果健康果
    LDA竖放柄朝上371098.440390.63
    竖放柄朝下37001002481.25
    横放柄朝输送方向380296.8811259.38
    横放柄垂直输送方向371492.1921062.50
    MD竖放柄朝上134093.754087.50
    竖放柄朝下142096.884087.50
    横放柄朝输送方向102096.884087.50
    横放柄垂直输送方向142096.884087.50
    KNN竖放柄朝上129085.942287.50
    竖放柄朝下1212081.253090.63
    横放柄朝输送方向911181.253381.25
    横放柄垂直输送方向912081.252287.50
    Table 2. Comparison of three discriminant modeling results of LDA, MD and KNN based on PCs extracted from transmission spectra (600~900 nm) of apples with four types of posture
    摆放姿态LVs训练集预测集
    错判个数准确率/%错判个数准确率/%
    病果健康果病果健康果
    竖放柄朝上76385.940293.75
    竖放柄朝下78185.941193.75
    横放柄朝输送方向72190.634290.63
    横放柄垂直输送方向212081.254087.50
    Table 3. PLS-DA results using transmission spectra (600~900 nm) of apples with 4 types of posture
    方法摆放姿态训练集预测集
    错判个数准确率
    /%
    错判个数准确率
    /%
    病果健康果病果健康果
    ELM竖放柄朝上6090.631290.63
    竖放柄朝下1787.502287.50
    横放柄朝输送方向1492.192190.63
    横放柄垂直输送方向2589.061290.63
    SVM竖放柄朝上0789.061096.88
    竖放柄朝下01281.250487.50
    横放柄朝输送方向00100.004184.38
    横放柄垂直输送方向1492.194184.38
    Table 4. Comparison of two modeling results of ELM and SVM using transmission spectra (600~900 nm) of apples with 4 types of posture
    摆放姿态LVs训练集预测集
    错判个数准确率/%错判个数准确率/%
    病果健康果病果健康果
    竖放柄朝上44093.750293.75
    竖放柄朝下313178.132093.75
    横放柄朝输送方向41295.311484.38
    横放柄垂直输送方向312179.694087.50
    Table 5. PLS-DA results using transmission spectra (690~720 nm) of apples with 4 types of posture
    Kai QIN, Gang CHEN, Jian-yi ZHANG, Xia-ping FU. Optimization of Fruit Pose and Modeling Method for Online Spectral Detection of Apple Moldy Core[J]. Spectroscopy and Spectral Analysis, 2021, 41(11): 3405
    Download Citation