Ground Plane Motion Camera Models

We show that it is possible to build an application-specific motion constraint into a general hierarchy of camera models so that the resulting algorithm is efficient and well conditioned; and the system degrades gracefully if the constraint is violated. The current work arose in the context of a specific application, namely smart convoying of vehicles on highways, and so ground plane motion (GPM) arises naturally. The algorithm for estimating motion from monocular, uncalibrated image sequences under ground plane motion involves the computation of three kinds of structure α,Β and γ for which algorithms are presented. Typical results on real data are shown.

[1]  Richard Szeliski,et al.  Recovering 3D Shape and Motion from Image Streams Using Nonlinear Least Squares , 1994, J. Vis. Commun. Image Represent..

[2]  Yoshiaki Shirai,et al.  Three-Dimensional Computer Vision , 1987, Symbolic Computation.

[3]  Michael Brady,et al.  Closing the loop on multiple motions , 1995, Proceedings of IEEE International Conference on Computer Vision.

[4]  Andrew Zisserman,et al.  Geometric invariance in computer vision , 1992 .

[5]  Long Quan,et al.  Relative 3D Reconstruction Using Multiple Uncalibrated Images , 1993, Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.