面向个性化制鞋,提出并实现了一种新颖的基于立体视觉原理的三维脚型建模与测量方法.该方法把标记点图案设置在袜子上,通过10台普通摄像机来采集用户穿着这种袜子的脚型图像;然后从脚型图像中提取出标记点并进行匹配,根据立体视觉原理重建出袜子上各个标记点的三维几何位置,并利用三角化处理以及曲面细分等手段获取用户的三维脚型;最后根据制鞋工艺要求测量出制鞋相关的参数.实验结果表明,该方法可瞬间完成无盲区图像的采集,有效地减少了生物体抖动所带来的误差,使误差在±1mm之内,可应用于个性化制鞋服务.
This paper presents a novel approach to scanning 3D foot models by stereo vision techniques based on the explicit markers woven in the socks. Ten cameras are carefully configured to make sure that each marker can be seen from at least two neighboring cameras. All markers are extracted by the modified Harris algorithm and their 3D positions are recovered in term of stereo vision principle. The 3D geometry of foot is acquired by the subdivision modeling based on the mesh composed of 3D marker points. Finally, the foot parameters are measured by the conventions for shoemaking. The experimental results show that the error is less than 1mm, which can well satisfy the requirement of customized shoemaking.