【作者】 龚光红； 宋子健； 杨帆； 王涛；

【Author】 GONG Guanghong SONG Zijian YANG Fan WANG Tao

【机构】 北京航空航天大学自动化科学与电气工程学院；

【摘要】 本文采用了基于多摄像机的光学测量技术捕获人体运动数据。通过可视化技术,生成了人体运动数据的二维三维检测工具。实现了大量数据的快速检测和修正。通过逆向运动学方法,人体运动的空间测量数据被转换成了关节角度数据,驱动三维高逼真度人体仿真模型。并基于原有数据基础对其他运动进行了建模仿真,效果良好。基于 Harris 算法和 Camshift 算法对人脸角点进行检测与追踪,并对虚拟人表情仿真进行了讨论。更多还原

【Abstract】 The optic measure method based on the multi camera was a- dopted to capture human motion data in this paper.By the visualization technique,the 2D and the 3D visual inspection tools for human motion data are created to implement the rapid inspection and correction of the large volume of data.The modeling method of hierarchy and articulating is used.By kinematics method,Motion angles of articulations are calcu- lated upon the spatial measured data,and are used to drive the three-di- mension high-resolution body mode].On basis of captured motion data, cut and try method was used to simulate the similar action of the cap- turod.There is a good result in application.Harris algorithm and Cam- shift algorithm are applied for face corner detection.,then face expression simulation of virtual human is discussed.更多还原