• Spectroscopy and Spectral Analysis
  • Vol. 41, Issue 12, 3844 (2021)
Ai-guo OUYANG*, Qi-ming WAN, Xiong LI, Zhi-yi XIONG, Shun WANG, and Qi-cheng LIAO
Author Affiliations
  • School of Mechatronics & Vehicle Engineering, East China Jiaotong University, Intelligent Electromechanical Equipment Innovation Research Institute, Nanchang 330013, China
  • show less
    DOI: 10.3964/j.issn.1000-0593(2021)12-3844-07 Cite this Article
    Ai-guo OUYANG, Qi-ming WAN, Xiong LI, Zhi-yi XIONG, Shun WANG, Qi-cheng LIAO. Research on Rich Borer Detection Methods Based on Hyperspectral Imaging Technology[J]. Spectroscopy and Spectral Analysis, 2021, 41(12): 3844 Copy Citation Text show less
    Rice borer holes of rice stem
    Fig. 1. Rice borer holes of rice stem
    Hyperspectral image acquisition system
    Fig. 2. Hyperspectral image acquisition system
    Spectral reflectance curves of 120 samples
    Fig. 3. Spectral reflectance curves of 120 samples
    The first five PC images
    Fig. 4. The first five PC images
    PC3 spectral weight coefficient
    Fig. 5. PC3 spectral weight coefficient
    RGB images at 668.8 and 750 nm
    Fig. 6. RGB images at 668.8 and 750 nm
    Flow chart of pest detection of rice borer based on principal component analysis characteristic band
    Fig. 7. Flow chart of pest detection of rice borer based on principal component analysis characteristic band
    Hyperspectral reflectance curves of a single sample
    Fig. 8. Hyperspectral reflectance curves of a single sample
    Optimal band iteration threshold segmentation image
    Fig. 9. Optimal band iteration threshold segmentation image
    The results of wormhole extraction based on iterative threshold
    Fig. 10. The results of wormhole extraction based on iterative threshold
    Flow chart of wormhole extraction method based on iterative threshold
    Fig. 11. Flow chart of wormhole extraction method based on iterative threshold
    组合序号可见光波长/混合距离近红外波段/混合距离组合波段/混合距离
    1689.9 nm/0.348753.5 nm/0.319689.9, 753.5 nm/0.667
    2682.9 nm/0.346753.5 nm/0.319682.9, 753.5 nm/0.665
    3689.9 nm/0.348750 nm/0.316689.9, 750 nm/0.664
    Table 1. Results of hybrid-distance combinations
    检测方法检测波段正常水稻
    茎秆(60个)
    检测正确
    率/%
    虫害水稻
    茎秆(60个)
    检测正确
    率/%
    整体检测
    正确率/%
    基于PCA特征波段检测668.8 nm处误判数591.7010095.8
    750 nm处误判数886.7010093.3
    基于迭代阈值的最优波段检测753.5 nm处误判数493.3010096.7
    Table 2. Results obtained by two detection methods
    Ai-guo OUYANG, Qi-ming WAN, Xiong LI, Zhi-yi XIONG, Shun WANG, Qi-cheng LIAO. Research on Rich Borer Detection Methods Based on Hyperspectral Imaging Technology[J]. Spectroscopy and Spectral Analysis, 2021, 41(12): 3844
    Download Citation