| 基于单视频图像序列的人体三维姿态重建 |
| |
| 引用本文： | 张晓,宋士华.基于单视频图像序列的人体三维姿态重建[J].计算机工程与设计,2008,29(6):1588-1591. |
| |
| 作者姓名： | 张晓 宋士华 |
| |
| 作者单位： | 九江学院,机械工程学院,江西,九江,332005 |
| |
| 摘 要： | 提出了至少存在一个深度值已知点的约束条件下,基于单视频图像序列重建人体三维姿态的方法.利用已知间距的平面点阵来标定获得摄像机参数,在透视投影模型下,根据单视频图像序列中人体关节点的二维数据,重建其三维信息.并将人体运动序列按照运动突变点划分为若干子序列,有效消除了二义性的干扰,较为精确的实现了人体三维姿态的重建.给出了该方法的实验过程及计算结果,验证了该算法的可行性和精度.
|
| 关 键 词： | 计算机视觉 人因工程 相机标定 三维姿态重建 数字化人体模型 |
| 文章编号： | 1000-7024(2008)06-1588-04 |
| 修稿时间： | 2007年3月28日 |
Reconstruction of 3D human posture based on monocular vision image sequences |
| |
| ZHANG Xiao,SONG Shi-hua.Reconstruction of 3D human posture based on monocular vision image sequences[J].Computer Engineering and Design,2008,29(6):1588-1591. |
| |
| Authors: | ZHANG Xiao SONG Shi-hua |
| |
| Affiliation: | ZHANG Xiao,SONG Shi-hua(School of Mechanical Engineering,Jiujiang University,Jiujiang 332005,China) |
| |
| Abstract: | For recovering 3D human posture from monocular vision image sequences,a novel approach is proposed by using the con-straints of which at least one depth value of a joint point has been known in each frame.The plane template model is used to simplify the calibration of CCD camera.And in the pin-hole perspective imaging model,3D characters of joint points could be obtained from 2D monocular vision image sequences.The adverse impacts from two feasible solutions are greatly eliminated by dividing human motion i... |
| |
| Keywords: | computer vision ergonomics camera calibration reconstruction of 3D human posture digital human model |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
