基于视觉的LED照明产品的自动装配

发布时间：2018-10-13 16:17

【摘要】：目前,LED照明整灯的装配作业大多由熟练的操作人员手工完成,或者由PLC程控半自动线完成。LED照明产品元器件的尺寸一般偏小,装配精度要求较高。手工装配对操作人员的专业技能要求很高,生产效率低,产品一致性差;随着市场对LED产品性能与产量需求的不断提高,手工装配已难以适应大批量生产,自动化装配在所必然。本文结合本公司新型LED整灯全自动装配流水线实际需求,研究了机器视觉在LED整灯自动装配线的应用。全文分别研究了工件在线的图像获得、识别、处理、应用,最终完成了LED新型灯具全自动生产线的研制、调试,最后交付生产工厂。本课题在研制开发的过程中,进行了以下研究:整理阐述了机器视觉理论的发展与应用现状及其对于自动化生产的意义;图像采集和处理方法;最小二乘法的直线与圆弧的拟合;机器人和视频图像的工件定位;系统的参数标定。根据机器视觉的工作原理,从整体上设计系统的相关布局,搭建了需要的识别系统工作舱。还以系统所具备的特点为基础,借助视觉软件的诸多算子,对相机外参数以及内参数进行科学合理的标定,此外还要标定机器人中的“手眼”系统。从此种标定算法来看,它的特征比较简单,而且具备较强的可行性特征,有着较高的精度。对静态分拣识别而言,探讨了动态工件的检测、分类定位,此法是按照触发时刻的特点空间位置,决定抓取时刻的工件所处的位置,属于空间向量平移方式,然后它借助图像处理软件所发出的相关指令,对机器人进行控制,从而实现工件分类。从这一套系统来看,它以视觉图像处理算子为基础,依赖于RC++开发平台而实现的,属于工件动态识别系统,而且在完成了识别和定位的调试后,获得了较高的准确率,完全能够满足事实上的生产要求。课题利用了基于视觉引导的机器人精密装配零件的先验知识,开发了相应的机器视觉自动引导软件,实际结果证明具有良好的工作效率和稳定性。自动装配线使用带有视频识别的机器人来替代人工加机器的半自动化装配和人工外观品质检测、码放,完成LED发光板、控制电路、灯头螺壳、外罩到LED整灯,从零部件组合\接插(镶嵌)\导线焊接\热熔封装\通电测试\商品喷码(字)\外观品质检测和托盘置放全部工序。通过视频系统、总线系统与机器手执行机构、链式传送机构、装备夹持平台有效而无干涉地协调运行,通过时间节拍的优化,使得生产效率达到前所未有的程度,设备综合效率达到85%,装配线2个人操作,每小时生产1800个单元,是传统人工生产线(每单元11人)的5倍。一次通过率也超过80%。达到了项目的开发设定目标,为行业开创了机器人和视频设别技术,推进生产设备自动化高效、优质途径。
[Abstract]:At present, the assembly of LED lighting whole lamp is mostly completed by skilled operators by hand, or by PLC program-controlled semi-automatic line. The dimension of components of LED lighting products is generally small, and the assembly precision is high. Manual assembly requires very high professional skills of operators, low production efficiency, poor product consistency, with the continuous improvement of market demand for the performance and output of LED products, manual assembly has been difficult to adapt to mass production. Automatic assembly is inevitable. In this paper, the application of machine vision in automatic assembly line of LED whole lamp is studied according to the actual requirement of automatic assembly line of new LED lamp. In this paper, the on-line image acquisition, recognition, processing and application of the workpiece are studied respectively. Finally, the development and debugging of the new LED automatic lamp production line are completed, and finally delivered to the production factory. In the course of the research and development of this subject, the following researches are carried out: the development and application of machine vision theory and its significance for automatic production, the method of image acquisition and processing, the development and application of machine vision theory are summarized. The fitting of the straight line and arc of the least square method, the localization of the robot and video image, and the calibration of the system parameters. According to the working principle of machine vision, the related layout of the system is designed as a whole, and the work module of the identification system is built. Based on the characteristics of the system and with the help of many operators of the visual software, the external and internal parameters of the camera are calibrated scientifically and reasonably. In addition, the hand-eye system in the robot is calibrated. From the point of view of this calibration algorithm, its features are relatively simple, and it has strong feasibility features and high accuracy. For static sorting recognition, the detection and classification of dynamic workpieces are discussed. This method is based on the characteristic spatial position of trigger time and determines the position of the workpiece at grab time, which belongs to the mode of translation of space vector. Then it controls the robot with the help of the relevant instructions issued by the image processing software and realizes the job classification. From the point of view of this system, it is based on the visual image processing operator and depends on the RC development platform. It belongs to the workpiece dynamic recognition system. Fully able to meet the actual production requirements. Based on the prior knowledge of robot precision assembly parts based on vision guidance, the corresponding machine vision automatic guidance software is developed. The practical results show that the software has good working efficiency and stability. The automatic assembly line uses a robot with video recognition to replace the semi-automatic assembly and manual appearance quality detection of the manual adding machine, code the LED, complete the LED light emitting board, control circuit, lamp head snail shell, outer cover to the LED whole lamp, From component combination\ insert (mosaic)\ wire welding\ hot melt package\ power test\ commodity spray code (word)\ appearance quality testing and pallet placement all processes. Through video system, bus system and machine hand actuator, chain transmission mechanism, equipment holding platform effectively and without interference coordinated operation, through the optimization of time beat, the production efficiency reached unprecedented degree. The overall efficiency of the equipment is 85 and the assembly line is operated by 2 individuals. The production of 1800 units per hour is 5 times that of the traditional manual production line (11 people per unit). The pass rate is more than 80% at a time. It has achieved the goal of project development, created robot and video technology for the industry, and promoted the production equipment automation, high efficiency and high quality way.
【学位授予单位】：上海交通大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP391.41;TM923.34

【相似文献】