基于线结构光扫描的足背三维轮廓重构  

Instep 3D Outline Reconstruction Based on Line-structured Light Scanning

作　　者：李新华[1] 袁振宇[1] 张涛[2,3] 姚志明[3] 

机构地区：[1]安徽大学计算智能与信号处理教育部重点实验室,合肥230039 [2]中国科学技术大学自动化系,合肥230027 [3]中国科学院合肥智能机械研究所,合肥230031

出　　处：《计算机工程》2014年第2期246-249,255,共5页Computer Engineering

基　　金：国家自然科学基金资助项目(61172127);中科院仪器研制基金资助重大项目(JCKY2012034)

摘　　要：针对人体足背的三维建模问题,提出一种基于线结构光扫描的新型重构及测量方案。采用2台高分辨率CMOS摄像头,从不同角度拍摄照射在足背上的线结构光线,通过灰度重心法细化图像中的结构光线,经坐标转换获得足背的点云数据。对足背点云数据后跟处的盲区采用最小二乘曲线拟合法进行修复,并运用三维Delaunay网格化算法实现曲面重构,获得完整的足背三维轮廓。实验结果表明,该方案程序运行稳定,足部主要参数的绝对误差均小于4 mm,具有较高的测量精确度。Aiming at the problem of human instep 3D modeling, a new scheme that uses line-structured light scanning to reconstruct 3D model and measure parameters is presented. Line structure light photos are taken by 2 high resolution CMOS cameras from different angles. To get the point cloud data of instep, the scanning images are thinned by gray-gravity method and image coordinates is transformed to world coordinates. The least squares curve fitting algorithm is used to fix the blind data in human heel zone. And the 3D Delaunay triangulation algorithm is used to restructure the curved face of instep and get a complete human instep 3D model. Experimental result shows that the running of the scheme is stable, the absolute error of the main parameters of foot is less than 4 mm, and it has high accuracy.

关 键 词：线结构光扫描 三维重构 足背点云 曲线拟合 曲面重构 参数测量 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]