基于三目视觉标定的机器人精确移载视觉系统

发布时间：2018-01-24 09:57

本文关键词： 三目视觉点标定最小均方误差匹配　出处：《组合机床与自动化加工技术》2017年11期 　论文类型：期刊论文

【摘要】：为了解决当前机器人视觉系统精度不稳定和单位作业时间过长,且缺乏自动补偿功能的不足,文章分别从系统硬件结构改进和软件算法优化的角度出发,提出了基于三目视觉的机器人视觉系统。首先,搭建PLC感应器、机器人和视觉PC之间的逻辑通信通道,达到材料运行状态实时获取与处理的目的。然后,根据机器人与视觉坐标系之间的缩放旋转平移参数,设计9点标定法,对三个工业相机进行标定,达到图像像素坐标与机器人世界坐标的精准绑定。最后,根据产品工程基准,选取感兴趣区域,优化最小均方误差匹配算法,达到系统精度与速度的平衡。实验测试结果显示:与当前机器人视觉系统相比,文中系统拥有更高的视觉准确度与作业效率。
[Abstract]:In order to solve the current robot vision system precision instability and unit operation time is too long, and the lack of automatic compensation function, this paper respectively from the point of view of system hardware structure improvement and software algorithm optimization. A robot vision system based on trinocular vision is proposed. Firstly, the logic communication channel between PLC sensor, robot and visual PC is built to obtain and process the material running state in real time. According to the zoom rotation and translation parameters between robot and visual coordinate system, a 9-point calibration method is designed to calibrate three industrial cameras to achieve the precise binding of image pixel coordinates and robot world coordinates. According to the product engineering datum, the region of interest is selected, and the minimum mean square error matching algorithm is optimized to achieve the balance between system precision and speed. The experimental results show that: compared with the current robot vision system. The system has higher visual accuracy and work efficiency.
【作者单位】：河南广播电视大学信息工程学院;
【基金】：河南省科技攻关项目(162102210122) 河南省科技成果转化项目(150201180010)
【分类号】：TP242
【正文快照】： 0引言近年来,机器视觉和机器人愈来愈多的应用到工业自动化生产,尤其是带有计算机视觉的机器人,为工厂无人化和智能化的实现与提高生产效率、质量,创造了有利条件[1-2]。目前,国内外开始把计算机的快速性、可靠性、结果得可重复性,与人类视觉的高度智能化和抽象能力相结合,产

【相似文献】