• Laser & Optoelectronics Progress
  • Vol. 56, Issue 16, 161008 (2019)
Wenhao Song, Bin Zhang*, Fengyu Li, Tengda Yang, Jianning Li, and Xiaohui Yang
Author Affiliations
  • College of Physical Engineering, Zhengzhou University, Zhengzhou, Henan 450001, China
  • show less
    DOI: 10.3788/LOP56.161008 Cite this Article Set citation alerts
    Wenhao Song, Bin Zhang, Fengyu Li, Tengda Yang, Jianning Li, Xiaohui Yang. Surface Crack Detection Algorithm for Nuclear Fuel Pellets[J]. Laser & Optoelectronics Progress, 2019, 56(16): 161008 Copy Citation Text show less
    Flow chart of proposed method
    Fig. 1. Flow chart of proposed method
    Pretreatment process. (a) Original image; (b) mask region; (c) region of interest
    Fig. 2. Pretreatment process. (a) Original image; (b) mask region; (c) region of interest
    Window scanning trace
    Fig. 3. Window scanning trace
    Structure of CrackCNN
    Fig. 4. Structure of CrackCNN
    Loss and accuracy of CrackCNN
    Fig. 5. Loss and accuracy of CrackCNN
    Beamlets at different scales. (a) Scale is 0; (b) scale is 1; (c) scale is 2; (d) scale is 3
    Fig. 6. Beamlets at different scales. (a) Scale is 0; (b) scale is 1; (c) scale is 2; (d) scale is 3
    Comparison of different threshold methods
    Fig. 7. Comparison of different threshold methods
    Detection results of three methods
    Fig. 8. Detection results of three methods
    LayerKernel shapeOutput channelStrideVariable
    Conv13×3×1321320
    Pool13×3×1-20
    Conv23×3×3248113872
    Pool23×3×1-20
    Conv35×5×4864176864
    Pool33×3×1-20
    Conv43×3×6480146160
    Pool43×3×1-20
    Fc18×8×80100-512100
    Fc21002-202
    Table 1. Parameter configurations of CrackCNN
    MethodPrecisionRecallF-measureTime /s
    Beamlet0.7010.7830.7413.6
    Doublethresholdand tensorvoting0.7400.8030.776.93
    Proposedmethod0.7750.8360.8042.41
    Table 2. Comparison of different methods
    Wenhao Song, Bin Zhang, Fengyu Li, Tengda Yang, Jianning Li, Xiaohui Yang. Surface Crack Detection Algorithm for Nuclear Fuel Pellets[J]. Laser & Optoelectronics Progress, 2019, 56(16): 161008
    Download Citation