Recording of multiple videos in a large-scale space for large-scale virtualized reality

A system for recording multiple videos in a very large-scale space, such as a soccer stadium, is described. It overcomes the problems of difficulty in recording synchronized multiple videos and in calibrating multiple cameras in a large-scale space. The system can record completely synchronized multiple videos with an identical timestamp for every frame captured at the same instant. Camera calibration in a large-scale space requires setting up an accurate coordinate system of the 3D world and defining many landmarks. The proposed system does this by combining a calibration board with a 3D laser-surveying instrument. A method for synthesizing a novel view from the multiple videos is also described, along with some experimental results. The technique deserbed in this paper is essential to realize a virtualized reality in a large-scale space.

[1]  Takeo Kanade,et al.  Shape reconstruction in projective grid space from large number of images , 1999, Proceedings. 1999 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No PR00149).

[2]  Takeo Kanade,et al.  Virtualized Reality: Constructing Virtual Worlds from Real Scenes , 1997, IEEE Multim..

[3]  Robert T. Collins,et al.  A space-sweep approach to true multi-image matching , 1996, Proceedings CVPR IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[4]  Yuichi Ohta,et al.  3D image display with motion parallax by camera matrix stereo , 1996, Proceedings of the Third IEEE International Conference on Multimedia Computing and Systems.

[5]  Richard Szeliski,et al.  Image mosaicing for tele-reality applications , 1994, Proceedings of 1994 IEEE Workshop on Applications of Computer Vision.

[6]  Ramesh C. Jain,et al.  An emerging medium: interactive three-dimensional digital video , 1996, Proceedings of the Third IEEE International Conference on Multimedia Computing and Systems.