The Research Of 3D Human Animation Based On Human Motion Tracking From Monocular Video Sequences

The generation of realistic 3D character animation is a popular and difficult issue in the area of computer graphics and virtual reality. The animation based on motion capture is realistic and reusable, and it is widely used in animation production. Hardware equipment like Vicon is commercially used to capture body movement in most cases, but it has some disadvantages. Players need to wear maillot and put reflective devices on joints, so the movability is limited. Furthermore, hardware equipment is expensive so animation producing cost is very high.This thesis is based on previous research. By combining vision technology and graphics technology, a new 3D human animation based on human motion tracking from monocular video sequences is proposed. The new method is easy-to-use, low cost and has better performance. The key points of the thesis are as below:1. Based on previous research results and by combining vision technology and Graphics Technology, a new 3D human animation based on human motion tracking from monocular video sequences is proposed and implemented.2. Under the assumption of the existence of 2D joint positions, using proportion of orthogonal to simulate camera shooting process can get unique variant S according to movement continuance. Then 3D relative coordinate figures of joints can be calculated according to the human body skeleton.3. According to H-anim standard, virtual topology tree of human joints can be established. Hence continuous animation can be got according to existed joint OLA angle.4. In theory, a prototype system can be got from framework. The experiments show that tracking and reconstructing results of our system are accurate and effective. This system can be applied to human animation producing.