Shape from Motion
In Sect. 2.6, we analyzed the optical flow induced by orthographic projection of planar surface motion. We gave an analytical solution of the 3D recovery equations in terms of invariants constructed from image characteristics. These invariants correspond to irreducible representations of SO(2)—the group of rotations of the image coordinate system. We also discussed the geometrical meanings of these invariants. In the following, we give an analytical solution of the 3D recovery equations in terms of the invariants constructed in Sect. 2.6. We also study adjacency conditions of optical flow and their implications. Finally, we present a scheme of motion detection that does not require point-to-point correspondences between different image frames.
KeywordsIntersection Line Image Plane Optical Flow Planar Surface Flow Parameter
Unable to display preview. Download preview PDF.
- J. J. Gibson: The Perception of the Visual World (Houghton Mifflin, Boston, MA 1950)Google Scholar
- J. J. Gibson: The Senses Considered as Perceptual Systems (Houghton Mifflin, Boston, MA 1966)Google Scholar
- J. J. Gibson: The Ecological Approach to Visual Perception (Houghton Mifflin, Boston, MA 1979)Google Scholar
- K. Prazdny: Determining the instantaneous direction of motion from optical flow generated by a curvilinear moving observer. Comput. Vision, Graphics Image Process. 17, 124–248 (1981)Google Scholar
- K. Kanatani, K. Yamada: Model-based determination of object position and orientation without matching. J. Infor. Process. 12(1), 1–8 (1988)Google Scholar
- K. Sugihara: Automatic construction of junction dictionaries and their exploitation for analysis for range data. Proc. 6th Intl. Joint Conf. Artificial Intelligence, August 1979, Tokyo, Japan, pp. 859–864Google Scholar
- I. Chakravarty, H. Freeman: Characteristic views as a basis for three-dimensional object representation. Proc. SPIE Conf. Robot Vision, 336, May 1982, Arlington, VA, pp. 37–45.Google Scholar
- C. Thorpe, S. Shafer: Correspondence in line drawings of multiple views. Proc. 8th Intl. Joint Conf. Artificial Intelligence, August 1983, Karlsruhe, FRG, pp. 959–965Google Scholar
- G. Fekete, L. S. Davis: Property spheres: A new representation for 3-D object recognition, Proc. IEEE Workshop on Computer Vision: Representation and Control, April-May 1984, Annapolis, MD, pp. 192–201Google Scholar
- R. Scott: Graphics and prediction from models. Proc. DARPA Image Understanding Workshop, October 1984, New Orleans, LA, pp. 98–106Google Scholar