基于机器视觉的LED TV智能屏检系统的设计与实现
|
摘要
为更好地助力企业转型升级,在自动化生产中引入机器视觉提升工业机器人加工质量与效率,探讨了一种以光学识别为基础的机器视觉方法用于LED TV的屏检技术,实现亮/暗点、横/竖线、BLOB(几何形态)等缺陷的检测;完成了智能屏检系统总体设计,搭建了机器视觉检测的硬件系统,并研制了智能屏检软件系统。对实现智能屏检涉及的光源选择、相机参数确定、屏检算法等关键技术进行了详尽地阐述,且针对不同检测目标和不同尺寸的屏幕进行了实测与分析。实践表明,本装备具有一定的实用性,其在实现生产线上工业机器人的柔性和智能化水平具有重要的应用价值与现实意义,并为机器人应用范围的拓展提供示范。
For the purpose of better promoting the enterprise transformation with industry 4. 0, the machine vision was introduced in automatic production to improve the production quality and efficiency of industrial robots. A method of machine vision based on optical recognition for screen inspection technology used in LED TV was discussed to realize the flaw detection such as bright spot, the horizontal and vertical line, BLOB(geometric shape) detection. The design of intelligent screen inspection system was finished, a hardware system for vision inspection was built, and the software system on intelligent screen inspection was developed. The key technologies of intelligent screen inspection were described in detail, including the selection of light source, the determination of camera parameters, the screen inspection algorithm and the intelligent screen test scheme, etc.,and the measurement and analysis of different inspection targets and different sizes of screen were carried out.The practice shows that the equipment has certain practicability, and it has important application value and practical significance in realizing the flexibility and intelligence level of industrial robot in production line, and can provide the demonstration for extending robot application scope.
For the purpose of better promoting the enterprise transformation with industry 4. 0, the machine vision was introduced in automatic production to improve the production quality and efficiency of industrial robots. A method of machine vision based on optical recognition for screen inspection technology used in LED TV was discussed to realize the flaw detection such as bright spot, the horizontal and vertical line, BLOB(geometric shape) detection. The design of intelligent screen inspection system was finished, a hardware system for vision inspection was built, and the software system on intelligent screen inspection was developed. The key technologies of intelligent screen inspection were described in detail, including the selection of light source, the determination of camera parameters, the screen inspection algorithm and the intelligent screen test scheme, etc.,and the measurement and analysis of different inspection targets and different sizes of screen were carried out.The practice shows that the equipment has certain practicability, and it has important application value and practical significance in realizing the flexibility and intelligence level of industrial robot in production line, and can provide the demonstration for extending robot application scope.
引文
[1]Malamas E N,Petrakis E G M,Zervakis M,et al. A survey on industrial vision systems, applications and tools[J]. Image&Vision Computing,2003,21(2):171-188.
[2]黄青丹,张亚茹,裴利强,等.基于图像识别的仪表自动检定系统[J].测控技术,2016,35(10):153-156.
[3]梁耀聪.基于机器视觉的显示屏缺陷检测系统研发[D].广州:广东工业大学,2013.
[4]郑喜凤,宋新丽,刘贵华,等.一种基于区域最大值的LED显示屏亮度特征数据提取方法[J].液晶与显示,2008,23(4):453-457.
[5]唐捷.TFT面板表面缺陷检测高速图像采集与处理方法研究[D].合肥:合肥工业大学,2016.
[6]张学武,徐立中,石爱业,等.基于视觉仿生机理的成像目标检测和识别方法及感知计算[J].传感技术学报,2010,23(12):1736-1743.
[7]朱光,朱学芳,张华坤.复杂背景下TFT-LCD表面缺陷检测系统的设计[J].电子测量与仪器学报,2011,25(12).
[8] Wu B W,Fang Y C. Human vision model in relation to characteristics of shapes for the Mach band effect[J]. Applied Optics,2015,54(28):E181.
[9] Bills R E,Koliopoulos C. Optical Defect Inspection System:US,20140268105[P].2014.
[10] Daisuke H,Koji K. Defect Inspection Device,Optical Member Manufacturing System,and Optical Display Device Production System:2015025703[P]. 2015.
[11] Mozina M,TomazeviěD,Pernus F,et al. Automated visual inspection of imprint quality of pharmaceutical tablets[J].Machine Vision and Applications,2013,24(1):63-73.
[12] Chiou Y C. Intelligent segmentation method for real-time defect inspection system[J]. Computers in Industry,2010,61(7):646-658.
[2]黄青丹,张亚茹,裴利强,等.基于图像识别的仪表自动检定系统[J].测控技术,2016,35(10):153-156.
[3]梁耀聪.基于机器视觉的显示屏缺陷检测系统研发[D].广州:广东工业大学,2013.
[4]郑喜凤,宋新丽,刘贵华,等.一种基于区域最大值的LED显示屏亮度特征数据提取方法[J].液晶与显示,2008,23(4):453-457.
[5]唐捷.TFT面板表面缺陷检测高速图像采集与处理方法研究[D].合肥:合肥工业大学,2016.
[6]张学武,徐立中,石爱业,等.基于视觉仿生机理的成像目标检测和识别方法及感知计算[J].传感技术学报,2010,23(12):1736-1743.
[7]朱光,朱学芳,张华坤.复杂背景下TFT-LCD表面缺陷检测系统的设计[J].电子测量与仪器学报,2011,25(12).
[8] Wu B W,Fang Y C. Human vision model in relation to characteristics of shapes for the Mach band effect[J]. Applied Optics,2015,54(28):E181.
[9] Bills R E,Koliopoulos C. Optical Defect Inspection System:US,20140268105[P].2014.
[10] Daisuke H,Koji K. Defect Inspection Device,Optical Member Manufacturing System,and Optical Display Device Production System:2015025703[P]. 2015.
[11] Mozina M,TomazeviěD,Pernus F,et al. Automated visual inspection of imprint quality of pharmaceutical tablets[J].Machine Vision and Applications,2013,24(1):63-73.
[12] Chiou Y C. Intelligent segmentation method for real-time defect inspection system[J]. Computers in Industry,2010,61(7):646-658.