全文下载: 202105014.pdf
文章编号: 1672-6987(2021)05-0087-07; DOI: 10.16351/j.1672-6987.2021.05.014
曹梦龙, 靳利文(青岛科技大学 自动化与电子工程学院,山东 青岛 266061)
摘要: 针对机器人手眼标定过程中人工标定时间长和标定误差大的问题,以双目测量为基础,设计了一种基于深度相机的机器人手眼标定方法。深度相机对标定物进行成像并建立点云模型,通过标定物的特征点视差及深度值反映摄像机的空间关系,同时导入机器人基坐标系下坐标,计算手眼关系变换矩阵,完成机器人手眼关系标定。实验表明:该方法标定的平面误差不高于040 mm,空间误差不高于046 mm,标定时间不超过5 min。验证了该方法的正确性和有效性,提高了工业机器人标定过程的作业效率。
关键词: 自标定; 双目视觉; 手眼标定; 点云模型
中图分类号: TP 391文献标志码: A
引用格式: 曹梦龙, 靳利文. 一种搭载深度相机的机器人手眼标定方法[J]. 青岛科技大学学报(自然科学版), 2021, 42(5): 87-93.
CAO Menglong, JIN Liwen. A calibration method of robot hand and eye equipped with depth camera[J]. Journal of Qingdao University of Science and Technology(Natural Science Edition), 2021, 42(5): 87-93.
A Calibration Method of Robot Hand and Eye Equipped with Depth Camera
CAO Menglong, JIN Liwen
(College of Automation and Electronic Engineering, Qingdao University of Science and Technology, Qingdao 266061, China)
Abstract: In view of the robot hand-eye calibration process of artificial time is long and the calibration of the problem of the great error. Based on the binocular measuring, this paper designs a robot hand-eye calibration method based on the depth of the camera. Depth of the calibration imaging camera and a point cloud model, By parallax of calibration feature points and depth value reflect the space relationship between the camera and import coordinate robot base coordinate system at the same time, the moment of hand-eye relation transforms. Experiments show that the calibration methods of plane error is less than 040 mm, space error is less than 046 mm, calibration time not more than 5 min. To verify the correctness and validity of the method, and improve the operational efficiency of the industrial robot calibration process.
Key words: self-calibration; binocular vision; hand-eye relationship; Point cloud model
收稿日期: 2020-11-30
基金项目: 山东省自然科学基金面上项目(ZR2020MF087).
作者简介: 曹梦龙(1971—),男,副教授.