基于机器视觉的工业机器人抓取技术研析.pdf

华中科技大学硕士学位论文 I 摘要目前在生产线上工作的搬运工业机器人大都是通过示教或预编程对其进行操作的, 这样操作目标物体的初始位姿和终止位姿都是严格限定的,机器人只是完成点到点的动作,这样一来,生产线的柔性就很差,满足不了柔性生产系统对物料输送系统的要求。在这种情况下,为保证搬运工业机器人顺利高效的完成工作任务,引入视觉识别和定位就显得很有必要。小批量多品种的生产模式是未来制造业的发展趋势。而示教和离线编程会花费大量的生产准备时间,生产效率的问题将尤为突出。本文研究的目的是对生产线上的工业机器人控制系统进行开发设计,将计算机视觉技术引入原有的搬运工业机器人领域,利用机器视觉技术获取工件及其周围环境的信息,识别出所要操作的目标工件,并能做出决策来引导工业机器人完成对工件的抓取和放置等操作,提高生产线上工业机器人的柔性和智能水平。本文提出了一种基于五轴工业机器人系统中加入计算机视觉的解决方案。主要的研究工作如下:提出了一种简单且准确度较高的实现单目摄像机标定的方法, ;对五轴机器人的运动学进行了分析并实现了目标定位;通过以太网实现了上下位机的通信;针对实验抓取的特定目标物体设计了手爪的夹持规划;根据机器人软件平台的特点,采用两种方案对控制软件进行开发设计:通过操作控制面板上特定按键实现视觉控制视觉控制和通过执行宏变量程序实现视觉控制。关键词五轴工业机器人,摄像机标定,位姿计算,上下位机通信,夹持规划华中科技大学硕士学位论文 II Abstract At present, the industrial r obots for carrying which are us ed in production generally utilize teaching or pre-programming mode. Unde r this ways, operation of initial object pose and terminate pose is limited strictly. Robots can plete point-to-point movements. So the flexibility of production line is unsatisf actory, can not to meet the flexile production system’s requirements for material conveying syst em. In this case, it is necessary to introduce visual identification and positi oning to ensure the smooth a nd efficient handing industrial robot task. The production mode of small batch but many varieties is the development trend of future manufacturing, and demonstration a nd off-line programming will cost a lot of time to prepare for production, the production efficien cy problem will be particularly outstanding. The purpose of this study is to design the line of industrial robot control system. By puter vision technology into the origin al hanging industrial robot field, using the machine vision technology acquire the info rmation of workpiece and its surrounding environment, identify the target workpiece to operate and make decisi on to guide industrial robots to finish the grab and place operation to im prove the flexibility and intelligence robots. This paper puts forward a solution to join computer vision in wh