The Semantic Environmental Model Of The Complex Virtual Scene And Character Animation

Motion planning and character animation of virtual characters in complex virtual sceneare important problems in virtual reality field. Interaction between virtual characters andvirtual environment needs show and explain virtual environment, which make virtualcharacters obtain information of virtual environment fast and supports efficient perception,wander and other behaviors of characters. Showing and explaining virtual environment needconstruct an environmental model, but research on environmental models are to plain virtualscene mostly not complex virtual scene, and motion types of virtual characters in virtualenvironment are few. So we need research characters of complex virtual scene and construct asuitable environmental model for it, to enhance abilities of perception to environmentalinformation and varieties of motion types in complex virtual scene.Based on geometric models of complex virtual scene, the extraction method for semanticinformation of the sampling points forming the height map in the geometric model of thecomplex virtual scene was proposed. By position relations of the obstacles and semanticinformation of the sampling points, the path points were ascertained on which characterscould stand. Furthermore, semantic navigation graph of the complex virtual scene was formedby extraction obstacle semantics and motion semantics of complex virtual scene. Height map,modified height map and semantic navigation graph made up of the semantic environmentalmodel of the complex virtual scene, which make full thinking of height, stride and behaviorpreference of the characters. Our model enhances the ability of perceiving the environmentalinformation and distinguishing obstacles automatically for virtual character, realizingautomatization of motion matching and variety of behaviors.Based on semantic environmental model of the complex virtual scene, two methods wereemploied to get motion clips and best path for characters in complex virtual scene. They weremethods by utilizing semantic navigation graph of the complex virtual scene and method byprecomputed search tree. Finally, motion clips of characters were put onto the best pathtogether to compose sequence of characters in complex virtual scene.