Euclidean 3D reconstruction from stereo sequences with variable focal lengths

  • Marc Pollefeys
  • Luc Van Gool
  • Theo Moons
Geometric Invariance
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1035)


A stereo rig can be calibrated using a calibration grid, but recent work demonstrated the possibility of auto-calibration. There remain two important limitations, however. First, the focal lengths of the cameras should remain fixed, thereby excluding zooming or focusing. Second, the stereo rig must not purely translate, which however is the most natural type of motion. This also implies that these methods collapse when the motion comes close to being a translation.

The paper extends the literature to allow changes in focal lengths (these may be independent for both cameras) and purely translational motions of the stereo rig. First, the principal points of both cameras are retrieved. Changes in focal lengths are then dealt with through weak calibration. Each position of the rig yields a projective reconstruction. The projective transformation between them allows to first retrieve affine structure which subsequently is upgraded to metric structure, following the general outline described in [12].

Rather than posing a problem to the method, rig translation allows further simplifications and is advantageous for robustness.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    M. Armstrong, A. Zisserman, and P. Beardsley, Euclidean structure from uncalibrated images, Proc. 5th BMVC, 1994.Google Scholar
  2. 2.
    R. Deriche, Z. Zhang, Q.-T. Luong, and O. Faugeras. Robust recovery of the epipolar geometry for an uncalibrated stereo rig. Proc.ECCV'94, pp. 567–576, Springer-Verlag, 1994.Google Scholar
  3. 3.
    F. Devernay and O. Faugeras, From Projective to Euclidean Reconstruction, IN-SIGHT meeting Leuven, 1995.Google Scholar
  4. 4.
    O. Faugeras, What can be seen in three dimensions with an uncalibrated stereo rig, Proc.ECCV'92, pp.321–334, 1992.Google Scholar
  5. 5.
    R. Hartley, Estimation of relative camera positions for uncalibrated cameras, Proc.ECCV'92, pp.579–587, 1992.Google Scholar
  6. 6.
    R. Hartley, Euclidean reconstruction from uncalibrated views, in: J.L. Mundy, A. Zisserman, and D. Forsyth (eds.), Applications of invariance in Computer Vision, Lecture Notes in Computer Science 825, pp. 237–256, Springer, 1994.Google Scholar
  7. 7.
    M. Li, Camera Calibration of a Head-Eye System for Active Vision, Proc. ECCV'94, pp. 543–554, Springer-Verlag, 1994.Google Scholar
  8. 8.
    Q.T. Luong and T. Vieville. Canonic representations for the geometries of multiple projective views. Proc. ECCV'94, pp. 589–597. Springer-Verlag, 1994.Google Scholar
  9. 9.
    T. Moons, L. Van Gool, M. Van Diest, and E. Pauwels, Affine reconstruction from perspective image pairs, in: J.L. Mundy, A. Zisserman, and D. Forsyth (eds.), Applications of Invariance in Computer Vision, Lecture Notes in Computer Science 825, pp. 297–316, Springer, 1994.Google Scholar
  10. 10.
    M. Pollefeys, L. Van Gool, and M. Proesmans, Euclidean 3D reconstruction from image sequences with variable focal lengths, Technical Report K.U.Leuven, E.S.A.T./MI2, 1995.Google Scholar
  11. 11.
    C. Rothwell, G. Csurka, and O.D. Faugeras, A comparison of projective reconstruction methods for pairs of views, Proc. ICCV'95,pp. 932–937, 1995.Google Scholar
  12. 12.
    A. Zisserman, P.A.Beardsley, and I.D. Reid, Metric calibration of a stereo rig. In Proc. Workshop on Visual Scene Representation, Boston, MA, June 1995.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1996

Authors and Affiliations

  • Marc Pollefeys
    • 1
  • Luc Van Gool
    • 1
  • Theo Moons
    • 1
  1. 1.Katholieke Universiteit Leuven, E.S.A.T./MI2LeuvenBelgium

Personalised recommendations