基于真实场景的大规模虚拟人群的快速生成方法

A Quickly Virtual Crowds Modeling Method Based on Real Scene

利用Poser的三维人体建模功能和3DS MAX强大的修改功能,结合MAXScript脚本语言,实现了对三维人体模型骨架的自动生成、匹配和蒙皮。通过对人物动作的建立与分解,实现了连续多个人体三维模型的生成与导出,并由此建立了一个包含不同人体及动作姿态的角色库。通过调用库中的角色,并将由图像信息获得的人群分布参数赋予这些虚拟个体人,实现了基于真实场景的虚拟人群快速生成。

A method of automatic production, matching and outer-surface covering for 3D human body is realized using 3D human body modeling functions in Poser and the modifying functions in 3DS MAX combined with the script language of MAXScript. Then continuous action frames of multi-human body models are generated and exported by creating and decomposition the character＇s action, and a liberally included these models is created simultaneously. Finally, a quick method of creating massive virtual crowds is realized by loading these characters from liberally and assigning the distribution parameters from image processing.