REAL-TIME VISUALIZATION METHOD FOR LARGE-SCALE CROWD SIMULATION

Animating virtual crowds with a large number of autonomous agents are effective to enhance realism in virtual reality applications. In this paper we present some preliminary results concerning a real-time visualization methods for large crowds reside in virtual environments. Making good use of LOD, we adopt three resolution models as detailed, less detailed geometry and imposter to represent virtual agent. At the same time we divide virtual crowd into crowd, group, and agent. Then we continue to divide agent into leader- agent and general-agent. We only have to certain which model should be adopted for leader-agent in one group not for every agent for less time complexity. We demonstrate our methods through the manifest the large number of human crowd in the virtual environment.