摘要
提出一种根据计算机视觉原理来获取人体表面信息的三维人体扫描仪方案,利用摄像机和人体之间的相对运动采集连续的图像序列,对序列图像进行预处理,从而获得各个角度的人体轮廓图像.然后根据轮廓图像的边缘点求取图像中各个高度的人体的侧影轮廓.最后通过一系列的人体的外侧影轮廓来逼近人体的原始形状,进而重建出人体表面三维信息.实验结果表明,本方案和算法是低成本、合理、有效的,根据本方案研制出来的原型系统的性能达到了预定的目标.
A scheme of 3D-human body scanner is proposed, which acquires the surface information of a human body based on the principle of computer vision. The continuous image sequence was captured via the relative movement between the human body and the camera. The human body contour images for various angles were obtained while the pre-processing being applied on the image sequence. The human body profiles for various heights were acquired, according to the coordinates of the various edge points on the contour images. A series of profiles were used to approximate the original human body figure, and the 3D surface of the human body was reconstructed. The results of the experiments show that the scheme and algorithm are characterized by low-cost, reasonable, and feasible. The performance of the prototype system based on the scheme achieves the scheduled level.
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2008年第5期24-26,34,共4页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家自然科学基金资助项目(69775022)
关键词
图像处理
人体扫描
三维重建
图像序列
计算机视觉
image processing
body scanning
3D reconstruction
image sequence
computer vision
