Research On Detection Technology Of Flat Panel Display Based On Machine Vision

Along with the popularization of interactive electronic product andhigh demand of flat panel display of the increasingly, so the detectionscreen also becomes especially important. The development of machinevision system based on provides feasible schemes to solve this situation.Machine vision will not have the eye fatigue, has a higher resolution thanthe human eye accuracy and speed, and has outstanding advantages.Combined with the actual engineering project: development ofdetection system of intelligent ammeter LCD screen display, some keytechniques of application of machine vision research in the flat paneldisplay testing system, build plan, system of image acquisition, imageprocessing algorithm theory and application. Among them, the imageacquisition, correction, image detection algorithm and itsimplementation will be the focus of research.Construction machine screen was based detection system. Detectionplatform hardware consists of OMRON FZ-SC color camera, FZ3-350, computervision controller; system software design, the graphical developmenttools of LabVIEW and NI for image processing and detection of TMAQ Vision，VBAI component development platform, powerful and complete functionlibrary, shortened and open extensible development cycle.In the realization of detection algorithms and techniques, theoriesand algorithms mainly research on the display defect detection, includingthe calibration of camera and image correction, image calibration,filtering, pattern matching, and proposes an improved image distortiondegree considering two kinds of distortion of the simple plane template.In the LabVIEW platform and visual detection of NI vision builder AIsoftware development environment, comprehensive implementation of systemdevelopment and the technology for detection of liquid crystal display defects, in order to achieve the key steps of calibration, inspection linedefect, the defect detection.Finally, combined with the actual project, according to thedetection requirements of smart meter, the paper will study the keytechnology and algorithm applied to detect meter LCD display performance.Having established the platform and the design of the test process,firstly the author apply the planar template method and geometriccorrection steps against the inclined angle and camera distortion. Afterthe image grayed, according to the standard template establishment,through edge detection, geometric matching algorithm preliminary matchesthe target contours and to determine the coordinates of the center of theregion of interest. Then the author use the normalized, two values,filtering and other image processing and LCD display the value of the twotemplate difference compared to achieve results. In the end, the experimentis validated.