摘要
文章提出一种基于双目视觉的通用并联机构(动平台为圆形)末端位姿识别方法。该方法依托双目视觉,利用阈值分割将动平台区域特征从机构图像背景中分离,并建立边缘搜索数学模型提取动平台边缘点,再借助最小二乘法拟合边缘点组成的椭圆以求取动平台中心点像素坐标,通过坐标转换关系并基于Matlab标定工具箱,求解得到其空间三维坐标。相对于安装传感器和利用激光跟踪仪,该方法简单直接且有效,为并联机构的位姿监测提供了新的可行性思路。
This paper proposes an end pose recognition method of universal parallel mechanism（rotational platform is round） based on binocular vision. The method relies on binocular vision, separate the dynamic platform region features from the background of the mechanism image by using threshold segmentation, extract the edge points of the moving platform by establishing an edge search mathematical model, and then fit the ellipse composed of the edge points by using a least-squares method to obtain the pixel coordinates of the center point of the moving platform, and obtain the spatial three-dimensional coordinates by solving the coordinates transformation relationship and using the Matlab calibration toolbox. Compared with the installation of sensors and the use of laser trackers, the method is simple, direct and effective, and it provides a new feasibility for the position monitoring of parallel mechanisms.
作者
李潇涵
侯雨雷
Li Xiaohan;Hou Yulei(ShiJiaZhuang No. 27 Middle School;School of Mechanical Engineering, Yanshan University)
出处
《电子技术（上海）》
2018年第4期13-15,共3页
Electronic Technology
基金
燕山大学研究生创新资助项目(2017XJSS006)
关键词
双目视觉
并联机构
阈值分割
最小二乘法
Binocular vision
parallel mechanism
threshold segmentation
least square method
