基于视觉的机器人球阀智能装配技术研究

发布时间：2018-06-02 10:38

本文选题：球阀零件装配 + 机器视觉　；参考：《太原科技大学》2014年硕士论文

【摘要】：球阀是生产生活中不可或缺的一类零件，其应用广、需求大，在国民经济中占有重要地位，装配是球阀生产的后置工序，在其生产制造中占有重要的地位。目前，国内企业在生产球阀时，仍由熟练工人运用传统人眼识别的方式对其进行装配。这种方法容易受到工人责任心、技术水平以及疲劳程度等不稳定因素的影响，其已不能满足现代生产制造自动化、智能化、高精度高效率的发展要求。为此探求良好技术开发平台下的自动装配技术成为必然。自机器视觉技术出现以来，其就被不断地应用于各类场合，并且取得了快速的发展。在工业生产中使用机器视觉系统即为工业视觉系统，按照其使用功能的不同，一般可分为检测、定位和装配三类。本文在球阀装配中采用机器视觉技术，开发出基于视觉的机器人球阀智能装配系统。系统以CCD相机对零件图像的采集作为智能装配的“感知”阶段；编写的系统控制处理程序作为智能装配的“决策”阶段；机械手爪抓取等作为智能装配的“执行”阶段。本文对基于视觉的机器人球阀智能装配系统进行了研究，整个系统由人机交互控制、装配工作台、视觉以及机器人夹持作业等四个主要单元组成。设计了由2台CCD相机及工业镜头组成的视觉子单元、由气动柔性通用机械手爪组成的夹持作业子单元等；同时编写了图像采集处理以及相机标定等系统程序。首先，视觉子单元采集目标零件图像，，并对图像进行小波去噪、增强、二值化、数学形态计算等在内的预处理；然后通过边缘检测、模板匹配等方法对目标零件进行识别定位；最后经过LabVIEW处理程序得出零件在机器人局部坐标系即标定坐标系中的位姿信息，进而控制柔性机械手爪完成抓取、装配作业。通过现场实验，验证了所研制的视觉装配系统不仅能够满足零件装配精度要求，而且提高了产品的合格率和生产率，降低了对操作人员的要求和工作强度。
[Abstract]:Ball valve is an indispensable part in production and life, it is widely used, it has a large demand, it occupies an important position in the national economy, assembly is the post-process of ball valve production, and it occupies an important position in its production and manufacture. At present, in the production of ball valves, domestic enterprises are still assembled by skilled workers using traditional human eye recognition. This method is easy to be influenced by the unstable factors such as worker's sense of responsibility, technical level and fatigue degree. It can not meet the requirements of modern production automation, intelligence, high precision and high efficiency. Therefore, it is inevitable to explore the automatic assembly technology under the good technology development platform. Since the emergence of machine vision technology, it has been continuously applied in various situations, and has made rapid development. The use of machine vision system in industrial production is called industrial vision system. According to the different functions of machine vision system, it can be divided into three types: detection, positioning and assembly. In this paper, a robot ball valve intelligent assembly system based on vision is developed by using machine vision technology in ball valve assembly. The system takes the acquisition of part image by CCD camera as the "perception" stage of intelligent assembly, the system control processing program as the "decision" stage of intelligent assembly, and the grasp of manipulator claw as the "execution" stage of intelligent assembly. In this paper, the intelligent assembly system of robot ball valve based on vision is studied. The whole system is composed of four main units: human-computer interaction control, assembly table, vision and robot clamping operation. A visual subunit composed of two CCD cameras and industrial lens is designed, which is composed of pneumatic flexible general manipulator claw, etc. The system programs such as image acquisition and processing and camera calibration are also written. Firstly, the image of target parts is collected by visual subunit, and the image is pre-processed by wavelet denoising, enhancement, binarization and mathematical form calculation, and then the target parts are identified and located by edge detection, template matching and so on. Finally, the position and attitude information of the parts in the local coordinate system of the robot is obtained through the LabVIEW processing program, and then the flexible manipulator claw is controlled to complete the grab and assembly operation. Through field experiments, it is verified that the developed visual assembly system can not only meet the requirements of assembly accuracy, but also improve the qualified rate and productivity of the products, and reduce the requirements and working intensity of the operators.
【学位授予单位】：太原科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP242;TH134

【参考文献】