Ground Plane Rectification by Tracking Moving Objects

Most outdoor visual surveillance scenes involve objects moving on a ground plane. We present a new, fully automated technique for both affine and metric rectification of this ground plane (up to a scale factor) by simply tracking moving objects. In particular, we derive the necessary constraints on the image plane to ground plane projective transformation by observing objects which move at constant (world) velocity for some part of their trajectory. No knowledge of camera parameters is assumed. We describe a hierarchy of possible solutions, depending on the nature of the motion trajectories observed. We also show how to automatically detect degenerate cases where 2D rectification is not possible. Useful applications of the various types of rectification are presented. Our experiments demonstrate all the possible solutions on a variety of scenes, as well as some of the applications made possible by rectification.

[1]  W. Eric L. Grimson,et al.  Learning Patterns of Activity Using Real-Time Tracking , 2000, IEEE Trans. Pattern Anal. Mach. Intell..

[2]  C. Jaynes Multi-view calibration from planar motion for video surveillance , 1999, Proceedings Second IEEE Workshop on Visual Surveillance (VS'99) (Cat. No.98-89223).

[3]  David Liebowitz,et al.  Camera Calibration and Reconstruction of Geometry from Images , 2001 .

[4]  Bernhard P. Wrobel,et al.  Multiple View Geometry in Computer Vision , 2001 .

[5]  Ian D. Reid,et al.  Single View Metrology , 2000, International Journal of Computer Vision.

[6]  Mubarak Shah,et al.  Tracking and Object Classification for Automated Surveillance , 2002, ECCV.

[7]  Andrew Zisserman,et al.  Metric rectification for perspective images of planes , 1998, Proceedings. 1998 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.98CB36231).

[8]  Hironobu Fujiyoshi,et al.  Moving target classification and tracking from real-time video , 1998, Proceedings Fourth IEEE Workshop on Applications of Computer Vision. WACV'98 (Cat. No.98EX201).

[9]  Pierre Gurdjos,et al.  About conditions for recovering the metric structures of perpendicular planes from the single ground plane to image homography , 2000, Proceedings 15th International Conference on Pattern Recognition. ICPR-2000.

[10]  James Orwell,et al.  Learning Surveillance Tracking Models for the Self-Calibrated Ground Plane , 2002, BMVC.

[11]  Ramakant Nevatia,et al.  Self-calibration of a camera from video of a walking human , 2002, Object recognition supported by user interaction for service robots.