摘要
针对基于单目视觉的单个圆位姿测量解的二义性问题,提出一种融合距离图像和彩色图像消除二义性的方法。该方法首先用RGB相机获取的高分辨率彩色图像计算圆位姿,获得圆位姿的一组歧义解。然后将彩色图像和飞行时间TOF(time-ofFlight)相机获取的距离图像融合,形成一个高分辨率的距离图像。最后,根据距离图像恢复圆在RGB相机下的三维信息,剔除圆位姿中的虚假解,从而得到圆的真实位姿。实验结果表明,该方法能够有效地解决基于单目视觉的单个圆位姿测量的二义性问题,且能够准确获得较高精度的圆位姿。
In order to solve the equivocation problem in monocular vision-based single circle pose measurement,we presented an approach to eliminate equivocation by fusing distance image and colour image.First,the method calculates the circle pose using the high-resolution colour image acquired by RGB camera and thus obtains a set of equivocal solutions of circle pose.Then,it fuses the colour image and distance image captured by TOF camera to form a high-resolution distance image.Finally,according to distance image it restores the 3D information of circle in RGB camera,culls the false solutions in circle pose,so as to get real pose of circle.Experimental results indicated that this method can effectively solve the equivocation problem of single circle pose measurement based on monocular vision,and can accurately attain circle pose with higher precision.
作者
陈欣
高隽
张旭东
张骏
Chen Xin;Gao Jun;Zhang Xudong;Zhang Jun(School of Computer and Information, Hefei University of Technology, Hefei 230009, Anhui, China)
出处
《计算机应用与软件》
CSCD
2016年第6期190-194,219,共6页
Computer Applications and Software
基金
国家自然科学基金项目(61273237
61271121)
安徽省自然科学基金项目(11040606M149)
关键词
圆
位姿测量
二义性
距离图像
彩色图像
融合
Circle
Pose measurement
Equivocation
Distance image
Colour image
Fusion