Sufficient image structure for 3-D motion and shape estimation
Abstract
We derive sufficient conditions on image structure that permits determination of 3-D motion parameters and depth from motion relative a rigid surface in front of the camera. We assume that only the first order spatio-temporal derivative or of the image is given and that the image intensity is continuously differentiable everywhere or that image contours are continuously differentiable. This means that only the component of the image motion field orthogonal to iso-intensity contours, the so called normal flow, can be measured. By defining a tangent line at a point as the line orthogonal to the gradient or normal the sufficiency condition on image structure can be stated as:
If each point (x,y) in the infinitely extended image plane is the intersection of at least 6 tangent lines, it is possible to compute unique 3-D motion and positive depth from first order spatio temporal derivatives, except for specific combinations of surface texture and depth.
The exceptions are specific texture patterns for any surface, for which the problem is inherently ambiguous, e.g. the so called “barber-pole”. These patterns have the property that there exist a relative motion to the surface such that the image flow field lines are aligned with the contours of the image.
Keywords
Optical Flow Motion Parameter Normal Flow Tangent Line Image StructureReferences
- 1.F. Bergholm. Motion from flow along contours: A note on robustness and ambiguous cases. International Journal of Computer Vision, 3:395–415, 1988.Google Scholar
- 2.F. Bergholm and S. Carlsson, A “theory” of optical flow, CVGIP-IU 53, 1991, 171–188.Google Scholar
- 3.S.Carlsson, Global structure of the velocity field of a moving plane, Proc. of IEEE Computer Society Workshop on Computer Vision, Miami Beach, Florida pp. 190–194, 1987Google Scholar
- 4.Z. Duric, A. Rosenfeld and L. Davies, “Egomotion analysis based on the Frenet-Serret motion model”, Proc. 4:th ICCV, pp 703–712, Berlin 1993Google Scholar
- 5.N. da Vitoria Lobo and J.K. Tsotsos, Using collinear points to compute egomotion and detect non rigidity, Proc. CVPR, 344–350. 1991Google Scholar
- 6.O.D. Faugeras and S. Maybank. Motion from point matches: Multiplicity of solutions. International Journal of Computer Vision, 4:225–246, 1990.Google Scholar
- 7.O. Faugeras and T. Papadopoulo, “A theory of motion fields of curves” International Journal of Computer Vision, 10:2, 125–156, 1993.Google Scholar
- 8.C. Fermüller and Y. Aloimonos, The role of fixation in visual motion analysis, International Journal of Computer Vision, 11:2, 165–186, 1993.Google Scholar
- 9.D.J. Heeger and A.D. Jepson, Subspace methods for recovering rigid motion I: Algorithm and implementation, International Journal of Computer Vision, 95–117, 1992.Google Scholar
- 10.E.C. Hildreth, Computations underlying the measurement of visual motion, Artificial Intelligence, 23, 309–354, 1984Google Scholar
- 11.B.K.P. Horn and B. Schunck. Determining optical flow. Artificial Intelligence, 17:185–203, 1981.Google Scholar
- 12.B.K.P. Horn. Motion fields are hardly ever ambiguous. International Journal of Computer Vision, 1:259–274, 1987.Google Scholar
- 13.Y. Liu and T.S. Huang. Estimation of rigid body motion using straight line correspondences. Computer Vision, Graphics, and Image Processing, 43:37–52, 1988.Google Scholar
- 14.H. C. Longuet-Higgins and K. Prazdny. The interpretation of a moving retinal image. Proceedings of the Royal Society, London B, 208:385–397, 1980.Google Scholar
- 15.H.H. Nagel. On the estimation of optical flow: Relations between different approaches and some new results. Artificial Intelligence, 33:459–483, 1987.Google Scholar
- 16.S. Negahdaripour and B.K.P. Horn, A direct method for locating the focus of expansion, Computer Vision, Graphics, and Image Processing, 46:303–326, 1989.Google Scholar
- 17.Philip J. ‘A Noniterative Algorithm for Determining all Essential Matrices Corresponding to Five Point Pairs', Technical report: TRITA-NA-P9320, Royal Institute of Technology, Stockholm, 1993Google Scholar
- 18.M. Spetsakis and J. Aloimonos. Structure from motion using line correspondences. International Journal of Computer Vision, 1:171–183, 1990.Google Scholar
- 19.O. Tretiak and L. Pastor, Velocity estimation from image sequences with second order differential operators, Proc. 7:th Int. Conf. Pattern Rec., Montreal, Canada, pp. 16–19, 1984Google Scholar
- 20.A. Verrri, F. Girosi and V. Torre, Differential techniques for optical flow, J. Opt. Soc. Am, A/Vol 7 No. 5 pp. 912–922, May 1990Google Scholar