|
Accession number;02A0234247
|
| Title;Position and Pose Estimation for Multiple Cameras in Distributed Human Tracking System. |
| Author;
YAMAZOE HIROTAKE
(Atrmediajohokagakuken)
UTSUMI AKIRA
(Atrmediajohokagakuken)
TETSUTANI NOBUJI
(Atrmediajohokagakuken)
YACHIDA MASAHIKO
(Osaka Univ. Graduate School of Engineering Sci., JPN)
|
Journal Title;Joho Shori Gakkai Kenkyu Hokoku
|
Journal Code:Z0031B
|
ISSN:0919-6072
|
|
VOL.2002;NO.2(CVIM-131);PAGE.117-124(2002)
|
| Figure&Table&Reference;FIG.8, REF.14 |
| Pub. Country;Japan |
| Language;Japanese |
| Abstract;We describe a camera calibration method to determine the positions and poses of all installed cameras in a system by using target human motions. In multiple-camera human tracking systems, increasing the number of cameras causes serious problems for camera calibration. Manual calibration becomes a heavy (boring) task for humans. In addition, maintaining the geometrical consistency among multiple cameras in a time sequence becomes difficult when the number of cameras is huge. To reduce these problems, we propose an automatic camera calibration method that enables both initialization and maintenance of all camera positions and poses. To avoid explosion of computation, our system maintains the calibration information (correlation and covariance matrices) in a distributed way. Each camera maintains a minimum amount of information. Experimental results show the effectiveness of our method. (author abst.) |
|
|
|
Related Articles;
|
