• Acta Optica Sinica
  • Vol. 38, Issue 8, 0815013 (2018)
Fengxia Guo*, Yun Wu*, Bin Li, and Jun Qi
Author Affiliations
  • Institute of Applied Technology, Heifei Institutes of Physical Science, Chinese Academy of Sciences, Hefei, Anhui 230088, China
  • show less
    DOI: 10.3788/AOS201838.0815013 Cite this Article Set citation alerts
    Fengxia Guo, Yun Wu, Bin Li, Jun Qi. Accurate Detection Method for Surface Indentation of Cold Stamping Valve[J]. Acta Optica Sinica, 2018, 38(8): 0815013 Copy Citation Text show less
    Acquisition result of valve surface indentation at normal lighting
    Fig. 1. Acquisition result of valve surface indentation at normal lighting
    (a) Schematic drawing of reflection stripe detection device; (b) diagram of distortion stripes
    Fig. 2. (a) Schematic drawing of reflection stripe detection device; (b) diagram of distortion stripes
    (a) Original gray image; (b) median filtered image
    Fig. 3. (a) Original gray image; (b) median filtered image
    (a) Original image; (b) shadow detail enhancement result
    Fig. 4. (a) Original image; (b) shadow detail enhancement result
    Result of valve foreground segmentation. (a) Before segmentation; (b) after segmentation
    Fig. 5. Result of valve foreground segmentation. (a) Before segmentation; (b) after segmentation
    Graph of gray variation in a line on the stripe image
    Fig. 6. Graph of gray variation in a line on the stripe image
    Extraction results of fringe-centerlines. (a) Image of gray stripe (the indentation is inside the red circle); (b) image of fringe-centerlines (the indentation is inside the red circle); (c) enlarged image of gray stripe (the indentation is inside the red circle); (d) enlarged image of fringe-centerlines (the indentation is inside the red circle)
    Fig. 7. Extraction results of fringe-centerlines. (a) Image of gray stripe (the indentation is inside the red circle); (b) image of fringe-centerlines (the indentation is inside the red circle); (c) enlarged image of gray stripe (the indentation is inside the red circle); (d) enlarged image of fringe-centerlines (the indentation is inside the red circle)
    Detection algorithm flow chart of valve indentation based on stripe image
    Fig. 8. Detection algorithm flow chart of valve indentation based on stripe image
    (a) Noise point (annotated by a red point) and (b) distortion stripe around the hole in the image of the fringe-centerlines
    Fig. 9. (a) Noise point (annotated by a red point) and (b) distortion stripe around the hole in the image of the fringe-centerlines
    Test results of indentation on valve surface (the indentation annotated by the red dot). (a) Gray image; (b) image of fringe-centerlines; (c) enlarged image of fringe-centerlines
    Fig. 10. Test results of indentation on valve surface (the indentation annotated by the red dot). (a) Gray image; (b) image of fringe-centerlines; (c) enlarged image of fringe-centerlines
    Fringe images of valve with (a) tiny indentation and (b) smudginess
    Fig. 11. Fringe images of valve with (a) tiny indentation and (b) smudginess
    Templateof the testvalveNumber ofthe indentationvalveNumber ofidentification ofvalve with indentationNumber ofthe falsedetectionNumber ofthe missingdetectionAccuracyrate /%Falserate /%Missingrate /%
    1605912972.53.3
    25052319662
    3404121973.32.5
    4404022963.35
    5606132957.53.3
    65049239546
    Table 1. Experimental results of the cold stamping valve