Object point tracking in photogrammetric measurement of human movement

Motion capture systems based on photogrammetric principles have gained wide r use over recent years for applications ranging from clinical gait analysi s to the creation of character animation for video games and the motion pic ture industry. Fast and robust determination of target trajectories is of c rucial importance to the provision of optimal performance in the recovery o f motion analysis parameters via multi-camera, real-time videometric system s. In this paper, the target path generation sequence is discussed, followe d by an account of the formulation of a kinematic model based on the alpha- beta-gamma filter/predictor. A brief description of the photogrammetric fea ture correspondence determination follows, together with an account of the 2D and 3D marker tracking procedures that were adopted. Finally, initial ex perimental tests results are summarised. These results indicate that the me thods developed an provide robust performance and an operational speed whic h is fast enough to allow real-time 3D target tracking at standard video fr ame rates and beyond.