Laser & Optoelectronics Progress, Volume. 58, Issue 20, 2015009(2021)

Circular Hole Pose Measurement Method Based on Binocular Vision Epipolar Compensation

Jinyou Chen1,2 and Zhiwei Guan1,3、*
Author Affiliations
  • 1School of Automobile and Transportation, Tianjin University of Technology and Education, Tianjin 300222, China
  • 2Department of Automotive Engineering, Hunan Financial & Industrial Vocational-Technical College, Hengyang, Hunan 421002, China;
  • 3School of Automobile and Rail Transit, Tianjin Sino-German University of Applied Sciences, Tianjin 300350, China
  • show less

    Circular holes are important features in automated manufacturing and assembly, and accurately estimating the pose of a hole can contribute to high-precision processes. Therefore, we propose a method for circular hole pose measurement based on epipolar constraints of binocular vision. First, to improve the accuracy of contour extraction, a contour purification method based on morphological addition is proposed. Then, the ordinate of the hole is compensated according to the epipolar constraint, and the optimal abscissa is determined via Gaussian fitting to achieve precise matching. Finally, the fitted circle in space is used to determine the hole pose. Experimental results show that the contour extracted using the method closely follows the real hole contour, and the measurement accuracy of the hole position and radius is 0.05 mm. Compared with no compensation, the measurement accuracy of the aperture and hole spacing increases by 31.56% and 34.07% after compensation, respectively. The method provides a strategy to improve the accuracy of hole pose measurement, and meets the actual application requirements of manufacturing and assembly.

    Tools

    Get Citation

    Copy Citation Text

    Jinyou Chen, Zhiwei Guan. Circular Hole Pose Measurement Method Based on Binocular Vision Epipolar Compensation[J]. Laser & Optoelectronics Progress, 2021, 58(20): 2015009

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category: Machine Vision

    Received: Nov. 2, 2020

    Accepted: Dec. 22, 2020

    Published Online: Oct. 14, 2021

    The Author Email: Guan Zhiwei (guanzhiw5511@163.com)

    DOI:10.3788/LOP202158.2015009

    Topics