The results of two experiments showed that the solution of the skeleton point solved the problem caused by data loss due to the Kinect optical occlusion. Finally, using the least squares polynomial fitting to compensate and correct the lost skeleton coordinate data, it realizes the smoothness and rationality of human skeleton animation. Then, the dynamic acceleration of skeleton motion and the spatial characteristics of static limb motion were calculated based on the data of adjacent effective skeleton frames before and after the collision. Firstly, the constraint characteristics of human skeleton motion were analyzed, and the maximum constraint table and Mesh Collider were established. Based on the analysis of the spatial characteristics of gymnastics and the movement principle of the human body, this paper proposes a dynamic and static two-dimensional regression compensation algorithm. Most of the existing research focuses on local motion but lacks consideration of the whole human skeleton. It is necessary to reduce data noise via specific algorithms. However, its optical characteristics are inevitably affected by illumination and occlusion.

Microsoft Kinect has been widely used due to its advantages of low price and high frame rate. The intelligent training and assessment of gymnastics movements require studying motion trajectory and reconstructing the character animation. Figure 8 also shows that ZMP tra- jectory has significant deviations nearby footstep positions, which could be explained with the limitations of MoCap system measure- ment accuracy. It means that the human gait was properly balanced, satisfying the criteria of static (GCoM) and dynamic (ZMP) stability, except the fi- nal part of the trajectory at the coordinate 2 m, where ZMP curve lies outside of the support area. Figure 8 demonstrates that both GCoM and ZMP trajectories are localized close to the footsteps and support polygons. The footsteps (i.e., par- ticular footprints) were approximated by ellipses taking into an ac- count RFoot, RToe and RToeEnd coordinates (according to Fig. The solid curve repre- sents the ZMP trajectory and the dashed curve represents the CoM projection trajectory on the ground plane. results of ZMP and GCoM trajectories calculations on the ground plane which were computed from human body dynamics (based on the simplified approximation of the human body as a single weight point) are shown in Fig.

Ipi mocap studio setup