A Six Point Solution for Structure and Motion
The paper has two main contributions: The first is a set of methods for computing structure and motion for m≥ 3 views of 6 points. It is shown that a geometric image error can be minimized over all views by a simple three parameter numerical optimization. Then, that an algebraic image error can be minimized over all views by computing the solution to a cubic in one variable. Finally, a minor point, is that this “quasi-linear” linear solution enables a more concise algorithm, than any given previously, for the reconstruction of 6 points in 3 views.
The second contribution is an m view n ≥ 6 point robust reconstruction algorithm which uses the 6 point method as a search engine. This extends the successful RANSAC based algorithms for 2-views and 3-views to m views. The algorithm can cope with missing data and mismatched data and may be used as an efficient initializer for bundle adjustment.
The new algorithms are evaluated on synthetic and real image sequences, and compared to optimal estimation results (bundle adjustment).
KeywordsImage Point Singular Vector Geometric Error Point Solution Bundle Adjustment
- 1.P. Beardsley, P. Torr, and A. Zisserman. 3D model acquisition from extended image sequences. In Proc. ECCV, LNCS 1064/1065, pages 683–695. Springer-Verlag, 1996.Google Scholar
- 2.S. Carlsson. Duality of reconstruction and positioning from projective views. In IEEE Workshop on Representation of Visual Scenes, Boston, 1995.Google Scholar
- 3.S. Carlsson and D. Weinshall. Dual computation of projective shape and camera positions from multiple images. IJCV, 1998. in Press.Google Scholar
- 5.A. W. Fitzgibbon and A. Zisserman. Automatic camera recovery for closed or open image sequences. In Proc. ECCV, pages 311–326. Springer-Verlag, Jun 1998.Google Scholar
- 6.R. I. Hartley. Euclidean reconstruction from uncalibrated views. In J.L. Mundy, A. Zisserman, and D. Forsyth, editors, Proc. 2nd European-US Workshop on In-variance, Azores, pages 187–202, 1993.Google Scholar
- 7.R. I. Hartley. Projective reconstruction and invariants from multiple images. IEEE T-PAMI, 16:1036–1041, October 1994.Google Scholar
- 8.R. I. Hartley. Multilinear relationships between coordinates of corresponding image points and lines. In Proceedings of the Sophus Lie Symposium, Nordfjordeid, Norway (not published yet), 1995.Google Scholar
- 10.A. Heyden. Projective structure and motion from image sequences using subspace methods. In Scandinavian Conference on Image Analysis, Lappenraanta, 1997.Google Scholar
- 11.D. Jacobs. Linear fitting with missing data: Applications to structure from motion and to characterizing intensity images. In Proc. CVPR, pages 206–212, 1997.Google Scholar
- 12.S. Mahamud and M. Hebert. Iterative_projective reconstruction from multiple views. In Proc. CVPR, 2000.Google Scholar
- 13.S. J. Maybank and A. Shashua. Ambiguity in reconstruction from images of six points. In Proc. ICCV, pages 703–708, 1998.Google Scholar
- 14.P. F. McLauchlan and D. W. Murray. A unifying framework for structure from motion recovery from image sequences. In Proc. ICCV, pages 314–320, 1995.Google Scholar
- 15.L. Quan. Invariants of 6 points from 3 uncalibrated images. In J. O. Eckland, editor, Proc. ECCV, pages 459–469. Springer-Verlag, 1994.Google Scholar
- 16.L. Quan, A. Heyden, and F. Kahl. Minimal projective reconstruction with missing data. In Proc. CVPR, pages 210–216, Jun 1999.Google Scholar
- 18.P. Sturm and W. Triggs. A factorization based algorithm for multi-image projective structure and motion. In Proc. ECCV, pages 709–720, 1996.Google Scholar
- 19.C. Tomasi and T. Kanade. Shape and motion from image streams under orthography: A factorization approach. IJCV, 9(2):137–154, Nov 1992.Google Scholar
- 22.P. H. S. Torr and A. Zisserman. Robust computation and parameterization of multiple view relations. In Proc. ICCV, pages 727–732, Jan 1998.Google Scholar
- 23.W. Triggs. Factorization methods for projective structure and motion. In Proc. CVPR, pages 845–851, 1996.Google Scholar