Subspace methods for recovering rigid motion I: algorithm and implementation
International Journal of Computer Vision
IEEE Transactions on Pattern Analysis and Machine Intelligence
Robot Vision
Theory of Reconstruction from Image Motion
Theory of Reconstruction from Image Motion
Catadioptric Omnidirectional Camera
CVPR '97 Proceedings of the 1997 Conference on Computer Vision and Pattern Recognition (CVPR '97)
Ego-Motion Estimation Using Optical Flow Fields Observed from Multiple Cameras
CVPR '97 Proceedings of the 1997 Conference on Computer Vision and Pattern Recognition (CVPR '97)
Ego-Motion and Omnidirectional Cameras
ICCV '98 Proceedings of the Sixth International Conference on Computer Vision
Discrete and Differential Two-View Constraints for General Imaging Systems
OMNIVIS '02 Proceedings of the Third Workshop on Omnidirectional Vision
| Hi-index | 0.00 |
Motivated by the full field of view of insect eyes and their fast and accurate estimation of egomotion, we constructed a system of cameras to take advantage of the full field of view (FOV) constraints that insects use. In this paper, we develop a new ego-motion algorithm for a rigidly mounted set of cameras undergoing arbitrary rigid motion. This egomotion algorithm combines the unambiguous components of the motion computed by each separate camera. We prove that the cyclotorsion is resistant to errors and show this empirically. We show how to calibrate the system with two novel algorithms, one using secondary cameras and one using self calibration. Given this system calibration, the new 3D motion algorithm first computes the rotation and then the 3D translation. We apply this algorithm to a camera system constructed with four rigidly mounted synchronized cameras pointing in various directions and present motion estimation results at www.cfar.umd.edu/~pbaker/argus.html.