Based On The Physics Of The Virtual Role Of Movement To Redirect Research And Realization

Been quick development in recent years, computer graphics has been widely used in many field such as special effects, 3D game manufacture, multimedia education and war game. Character animation which has been combined with mathematics, physics and computer science is an important research topic. Although the technology of character animation is relatively mature, this field, as the branch of computer graphics, still faces a grand challenge and opportunity. In future, the animation effect which we need is the virtual scene cannot be told apart by our naked eyes. So, it is a long way to go.Research on the editing of motion capture data is a hot spot issue in character animation. Motion retargeting technology can enrich the database of motion capture, increase the efficiency of character animation making and reduce the cost. By analyzing the main technologies in motion retargeting and combining with the national 863 project- Integrated Environment Supporting Digital Media Content Creation, this thesis focus on the following three aspects.1. Propose an approach focus on skeleton model with difference arbitrary topologies. Compared with the research on motion retargeting based on the same skeleton topologies, there is little research on motion retargeting aiming at quite different skeleton topologies. Base on human skeleton model, this thesis adopted constraint method and motion data mapping which considers the main joints that has great effect on the character motion to rectify the result of motion retargeting and keep the features of source motion maximally;2. Construct fairly perfect physical constraints. Only using the kinematic methods, modeling will be easy but the synthetic motion maybe lack of physical realistic; only using dynamic methods, the synthetic motion maybe satisfy the physical realistic requirement but the solving progress is complex and time consuming. This thesis adopted inverse kinematics method to solve the motion satisfying the geometrical constraints firstly, and then construct resolver of balance constraint, torque constraint and movement constraint. Input the source motion data and require the target motion data. Treating the physical constraints as the post-processing approach, this resolver only includes physical constraints. So, this approach reduced the modeling complexity and can make sure the synthetic motion keep the physical features;3. Combing with key-frame technology to synthesis the new motion. This thesis adopted key-frame technology as an important supplement. We have synthesized new motion and make the motion transition more flexible by using key-frame method.Lots of motion capture data files have been used as the source files. The experiment has testified that the algorithm-motion retargeting based on physics can generate realistic and slick animation which complies with the law of physics and the skeleton topology of the target character. At the meantime, this new animation can keep the characteristics of the source animation at maximum degree.