摘要
针对电池PACK线电池镍片摆放难以自动化问题,以某电池自动化生产线中机器人对电池镍片的识别与无序抓取为例,进行了理论分析和测试试验研究。用Eye-to-Hand的三点手眼标定法建立了机器人与视觉相机之间的仿射变换矩阵;利用Halcon软件平台对拍摄电池镍片图像进行零件特征提取和基于形状的模板匹配算法进行镍片的识别,并研究了镍片各种位姿情况下对识别速度、准确性的影响;最后搭建了测试平台进行了验证。研究结果表明,该方法的识别速度达到了32.3 ms,识别准确度达到了0.9521,一次性可以识别10个以上不同位姿形态的镍片,可以精准有效地引导机器人对物料进行操作,满足了该自动化生产线的需求,对类似作业类型有较好的参考价值。
Aiming at the difficulty of automating the placement of nickel sheets in battery PACK line,the identification and disordered grabbing of nickel sheets in a battery automatic production line were taken as an example,and theoretical analysis and test were conducted.The affine transformation matrix between the robot and the visual camera is established by using the Eye-to-Hand three-point hand-eye calibration method.The feature extraction and shape-based template matching algorithm were used to identify the battery nickel slice images by Halcon software platform,and the influence of various positions and poses on the recognition speed and accuracy was studied.Finally,a test platform was built for verification.The research results show that the recognition speed of this method reaches 32.3 ms,and the recognition accuracy reaches 0.9521.It can identify more than 10 nickel pieces with different poses and shapes at one time,and can accurately and effectively guide the robot to operate the materials,which meets the needs of the automatic production line and has good reference value for similar types of operations.
作者
张烜
刘明尧
ZHANG Xuan;LIU Mingyao(School of Mechanical and Electrical Engineering,Wuhan University of Technology,Wuhan 430000,China)
出处
《自动化与仪表》
2024年第6期72-75,80,共5页
Automation & Instrumentation
关键词
手眼标定
模板匹配
目标定位
零件抓取
hand-eye calibration
template matching
target location
parts grabbin
