• DocumentCode
    3247474
  • Title

    Robocup field distortion calibration with variational image registration

  • Author

    Li, Guan ; Chengguo, Guo ; Rong, Xiong

  • Author_Institution
    Nat. Lab. of Ind. Control Technol., Zhejiang Univ., Hangzhou, China
  • fYear
    2004
  • fDate
    20-22 Oct. 2004
  • Firstpage
    338
  • Lastpage
    341
  • Abstract
    The Global Vision System is widely used among teams of Robocup F180 (Small Size League). This architecture as well as the 1-hour preparation time limit before each match poses strict requirements to the accuracy and efficiency of the calibration method against pitch field distortions. Because of the curvature-preserving feature of these distortions, we propose a brand-new image registration approach, which does not require calibrating the camera, but directly maps the distorted image to the real pitch. While this method substantially simplifies the pre-match manipulation, it also maintains high accuracy - our experimental result shows that, with double camera, the mean error ratio is limited within 0.22%, accurate enough for Robocup matches.
  • Keywords
    calibration; distortion; image registration; robot vision; variational techniques; Global Vision System; Robocup F180; curvature-preserving feature; double camera; field distortion calibration; image registration; pitch field distortions; variational image registration; Calibration; Cameras; Computer aided instruction; Humans; Image registration; Industrial control; Machine vision; Robot vision systems; Robustness; Tires;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Intelligent Multimedia, Video and Speech Processing, 2004. Proceedings of 2004 International Symposium on
  • Print_ISBN
    0-7803-8687-6
  • Type

    conf

  • DOI
    10.1109/ISIMP.2004.1434069
  • Filename
    1434069