A statistical approach to structure and motion from image features
Poster Papers
First Online:
Abstract
The estimation of structure and motion from image sequences using corresponding points, lines, conics and structured patches is treated. Recent research has provided good tools for obtaining good initial estimates of structure and motion using point, line, conic and curve correspondences. These estimates are, however, not so accurate. In this paper it is shown how to obtain statistically optimal estimates of structure and motion using a combination of such image feature correspondences. The question of using proper weighting is important when different types of features are combined. We show how weights can be chosen in a statistical optimal sense. Experiments with real data are used to evaluate every step of the algorithm.
Keywords
Bundle adjustment Points Lines Conics Patches Download
to read the full conference paper text
References
- 1.K. Åström and A. Heyden. Stochastic analysis of sub-pixel edge detection. In Proc. International Conference on Pattern Recognition, Vienna, Austria, pages 86–90, 1996.Google Scholar
- 2.K. Åström and A. Heyden. Stochastic analysis of image acquisition and scalespace smoothing. In J. Sporring, M. Nielsen, L. Florack, and P. Johansen, editors, Gaussian Scale-Space Theory. Kluwer Academic Publishers, 1997.Google Scholar
- 3.K. B. Atkinson. Close Range Photogrammetry and Machine Vision. Whittles Publishing, 1996.Google Scholar
- 4.R. Berthilsson and K. Åström. Reconstruction of 3d-curves from 2d-images using affine shape methods for curves. In Proc. Conf. Computer Vision and Pattern Recognition, 1997.Google Scholar
- 5.R. Berthilsson, K. Åström, and A. Heyden. Projective reconstruction of 3d-curves from its 2d-images using error models and bundle adjustments. In Proc. 10th Scandinavian Conference on Image Analysis, pages 581–588, 1997.Google Scholar
- 6.O. D. Faugeras and B. Mourrain. On the geometry and algebra on the point and line correspondences between n images. In Proc. 5th Int. Conf, on Computer Vision, MIT, Boston, MA, pages 951–956. IEEE Computer Society Press, 1995.Google Scholar
- 7.R. I. Hartley. A linear method for reconstruction from lines and points. In Proc. 5th Int. Conf. on Computer Vision, MIT, Boston, MA, pages 882–887, 1995.Google Scholar
- 8.R. I. Hartley. Lines and points in three views and the trifocal tensor. Int. Journal of Computer Vision, 22(2):125–140, 1997.CrossRefGoogle Scholar
- 9.F. Kahl and A. Heyden. Structure and motion from points, lines and conics with affine cameras. In Proc. 5th European Conf. on Computer Vision, Freiburg, Germany, 1997.Google Scholar
- 10.F. Kahl and A. Heyden. Using conic correspondences in two images to estimate the epipolar geometry. In Proc. 6th Int. Conf. on Computer Vision, Mumbai, India, pages 761–766, 1998.Google Scholar
- 11.S.D. Ma. Conic-based stereo, motion estimation, and pose determination. Int. Journal of Computer Vision, 10(1):7–25, 1993.CrossRefGoogle Scholar
- 12.L. Quan. Conic reconstruction and correspondence from two views. IEEE Trans. Pattern Analysis and Machine Intelligence, 18(2):151–160, February 1996.CrossRefGoogle Scholar
- 13.L. Quan and T. Kanade. Affine structure from line correspondences with uncalibrated affine cameras. IEEE Trans. Pattern Analysis and Machine Intelligence, 19(8), August 1997.Google Scholar
- 14.C. R. Rao. Linear Statistical Inference and Its Application. John Wiley & Sons, Inc., 1965.Google Scholar
- 15.G. Sparr. Simultaneous reconstruction of scene structure and camera locations from uncalibrated image sequences. In Proc. International Conference on Pattern Recognition, Vienna, Austria, pages 328–333, 1996.Google Scholar
- 16.M. E. Spetsakis and J. Y. Aloimonos. A unified theory of strucure and motion. In Proc. DARPA Image Understanding Workshop, 1990.Google Scholar
- 17.J. Weng, T.S. Huang, and N. Ahuja. Motion and structure from line correspondances: Closed-form solution, uniqueness, and optimization. IEEE Trans. Pattern Analysis and Machine Intelligence, 14(3), 1992.Google Scholar
Copyright information
© Springer-Verlag Berlin Heidelberg 1998