• DocumentCode
    2114969
  • Title

    A linear algorithm to estimate the plane of infinity

  • Author

    Shen, Fei ; Wang, Han

  • Author_Institution
    Sch. of Electr. & Electron. Eng., Nanyang Technol. Univ., Singapore
  • Volume
    3
  • fYear
    2002
  • fDate
    2-5 Dec. 2002
  • Firstpage
    1326
  • Abstract
    This paper considers the 3D reconstruction problem from an image sequence taken by a camera with constant internal parameters. Stratified method starts from a projective reconstruction, refines it into affine reconstruction and finally upgrades it to metric reconstruction. It was reported that the most difficult step is to obtain the affine reconstruction, that is, to find the plane at infinity in the projective coordinate frame. In this paper, we give a simple linear solution to estimate the location of infinity plane when the camera´s rotation is small. The developed method depends on the linear relationship among three main diagonal elements of the infinity homography of such two views. Three images are sufficient to give one unique solution, which can be used as the start point of nonlinear constraints. Simulations showed that out algorithm gives very close estimation to the exact location of the plane of infinity.
  • Keywords
    computer vision; image restoration; image sequences; video cameras; 3D reconstruction problem; affine reconstruction; camera; constant internal parameters; image sequence; infinity homography; infinity plane; linear algorithm; linear solutions; metric reconstruction; nonlinear constraints; projective coordinate frame; projective reconstruction; Calibration; Cameras; Computer simulation; Computer vision; Focusing; H infinity control; Image reconstruction; Image sequences; Layout; Nonlinear equations;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Control, Automation, Robotics and Vision, 2002. ICARCV 2002. 7th International Conference on
  • Print_ISBN
    981-04-8364-3
  • Type

    conf

  • DOI
    10.1109/ICARCV.2002.1234965
  • Filename
    1234965