The Design And Development Of LCD Line Defect Detection System

LCD (liquid crystal display) as the main display of all kinds of instruments andappliances has achieved fast development and mass production in the last ten years.However, a variety of defects will be caused in production process by the complexity ofLCD preparation process, so the defect detection naturally becomes an indispensablelink of LCD production process. At present, the liquid crystal line defect detection reliesmainly on labor, not only inefficient, but also prone to false alarms and undetected cases.This situation will undoubtedly greatly increase the cost of production. In order toimprove production efficiency, to enhance detection reliability, to reduce productioncosts, a set of LCD line automatic defect detection system is designed using machinevision approach.According to the material characteristics of LCD liquid crystal device and thedetection requirements of various defects such as short circuit and open circuit andimpurities in LCD screen line, the system solutions of LCD line defect detectionequipment is designed. Through the motor motion precise controlled by motion controlmodule, the camera image can be obtained by image acquisition module, then thestitching, cutting, processing and defect detection of acquisition image are completed inthe image processing module.According to the software design requirements of LCD line defect detection device,a full range of testing software solutions is designed, including basic information set,motion capture platform, the standard template image creation and defect detectionapplications four parts. Finally, based on the hardware platform of LCD line defectdetection device, the LCD line defect detection software written by C++and Qt istested under Visual Studio2008operating environment. Experimental results show that:the clear image of LCD line can be captured and the stitching of captured images can becompleted, then three lines defects including short circuit, open circuit and impuritiescan be accurately judged, following the outputs of "Defect List", including defect types,each defects coordinates corresponding to the original image, image read time and thetotal detecting time of current image and other information.