The Research Of Character Animation Editing Technique Based On Data And Motion Platform Implementation

With the rapid development of computer animation technology, character animation has been widely applied in virtual reality, entertainment, video games, education and training, sports and military simulation. Motion capture technology and the emergence of large-scale commercial applications greatly motivate the creation of animation for the human body and also overcome the shortcomings of the traditional key-frame animation methods. Therefore, it is necessary to study how to make full use of motion capture data and technology with traditional key frame animation for character animation.According to the contents involved in character animation creation, this article focuses on the following research areas:Firstly, this thesis show the description and analysis of the inverse kinematics methods based on Jacobi matrix and on this basis how to achieve a pseudo-inverse Jacobi matrix method and damped least squares method. Both of these methods are used to achieve some simple IK chains and full-body IK for virtual character.Secondly, in order to overcome the shortcomings caused by numerical methods which are too much manual intervention, unnatural results and so on, we present a combination of motion capture data and numerical iterative method for human motion editing. The main idea of this method is to classify locomotion into different types based on some particular type of movement data. So the restrictions are more close to the real situation and the results will be more natural. In addition, these constraints also have a different priority levels to better solve the conflicts between them. Experiments show that our proposed method is simple to use and can ensure the efficiency of real-time interaction. It can be easily carried out for different types of movement to adjust the details to create a new movement.Thirdly, the linear and nonlinear interpolation methods are used to fulfill motion editing based on keyframe technique. This thesis offers a variety ways to generate keyframe pose. You can choose FK/IK to generate a pose by hand, or pose key frames according to the frames coming from motion capture data, and also generate a pose through modifying existing mocap data. Experiments show that these algorithms can achieve a good motion sequence from corresponding key frames.Finally, this thesis implement a simple keyframe-based animation platform. This platform has good basic functionality and flexible scalability. It can be used to fast implement and verify new animation algorithm. Using this animation platform, we are able to create, edit and also set keyframe motion for ordinary geometry as well as virtual character. In addition, the platform also uses the XML Schema technology to save and load the objects as well as their animation information. This function makes the content created by this platform can be used by other programs and expand the scope of this application.Experiments show that the goal-oriented movement synthesis method which is based on motion capture data can effectively solve the problems caused by traditional inverse kinematics and it is very easy to set keyframe and also can guarantee the results to be more natural and reliable. We achieve the motion editing platform that can quickly and effectively create and edit keyframe-based animation, including the scene management, animation production and other functions.