Complex Character Animation Environment Automated Selection And Synthesis Of Implementation

Along with the continuous development of the computer hardware and the computer graphics algorithm, computer animation is becoming the hottest focus of current research. In the field of computer animation, character animation is a popular research direction. It has achieved great success in the field of digital entertainment, but there are still a lot of problems need to be solved.Character animation is the animating of characters in really environment, ensuring the reality of characters is the only way to make human motion real and increase the immersion. However, a character has multiple degrees of freedom, and its limb movements are complex. Besides, the figure is irregular, and movement style is protean. Human are familiar with their own movements, and they can easily detect the uncoordinated movement, even a walking posture can give enough information to identify an acquaintance. So, keeping the reality of the human motion is an important problem , also the first problem we focus on in this thesis.Research shows that character animation methods based on motion data can ensure the reality of character, because the data come from the real world activities of the performers. So, the motion data method is used to drive the character. However, the principle of the method is a simple recording and playback of three dimensional movement, it cannot provide more interactive controlling. This makes the data cannot be reuse and difficult to adapt to the application of strong interactions.In order to solve the problem, the motion data is built into a motion graph, so target motion generating is changed into graph search problem. This method makes every frame of motion data a node of motion graph, and create transitions between similar postures as edges of the graph. After that, linear structure of the original motion data is reconstructed as more flexibility graphic structure. However, the motion graph method has its inherent fault which the efficiency and functional of generated movements subject to its original motion capture database. In order to solve the defects of motion graph method, the motion graph is expanded and compressed. The purpose of expansion is increasing the function, and that of compression is reducing the capacity of the graph.The current character animation techniques rely on experienced animators, and is a heavy physical and mental work. Increasing the efficiency of character animation and reducing the participation of animators can effectively reduce the time and character animation production costs. However, due to its complexity and uncertainty, improving the animation production efficiency is a challenge. This is the second problem which is solved in this thesis.The usual application of character animation is producing animation of certain environment. In order to improve efficiency of character animation, a method to automatic synthesis character animation in complex environment is proposed.