基于多传感器融合的堆叠棒料识别和抓取  

Identification and Grasping of Stacked Rod Based on Multi-sensor Fusion

作　　者：卢欣龙 赵岩 代凡 李棋 杨凌云 方辉[1] LU Xinlong;ZHAO Yan;DAI Fan;LI Qi;YANG Lingyun;FANG Hui(Sichuan University,Chengdu 610000,China;Sichuan Dekepu CNC Machine Tool Co.,Ltd.,Suining 629000,China)

机构地区：[1]四川大学,成都610000 [2]四川省德克普数控机床有限公司,四川遂宁629000

出　　处：《机械工程师》2025年第3期27-31,35,共6页Mechanical Engineer

基　　金：2021年四川大学-遂宁市校地战略合作“揭榜挂帅”科技项目(2021CDSN-15)。

摘　　要：针对合金棒料等刀具原材料的抓取问题,提出了一种基于多传感器融合的方式完成识别和抓取棒料工作。采用眼在手外的安装方式并在完成手眼标定后,使用物体边缘轮廓等方法,得到了棒料抓取点和棒料方向角。通过移动激光测距仪,处理得到的距离数据,获取棒料中心位置的深度信息。最后引导机械臂到达指定位置进行棒料抓取。实验表明,在误差允许范围内,通过多传感器引导机械臂寻找并抓取棒料具有良好的准确性。This paper proposes a method for identifying and grasping rod materials such as alloy bars based on multi-sensor fusion.Employing an exterior setup of sensors relative to the manipulator and performing hand-eye calibration,this paper utilizes techniques like object edge contour analysis to determine the gripping points and orientation angle of the rod materials.Utilizing a moving laser rangefinder,depth information regarding the central position of the rod materials is obtained from processed distance data.Finally,the robotic arm is guided to the designated position for the gripping of rod materials.Experiments show that the robotic arm guided by multi-sensors to find and grasp bars has good accuracy within the error tolerance.

关 键 词：相机标定 多传感器融合 轮廓检测 物体抓取 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置] TH164[自动化与计算机技术—控制科学与工程]