LED Panel Intelligent Detecting System Based On Machine Vision

Large screen display has been widely applied in various areas, such as commercial conferences, teaching seminars, advertisement, industrial control and so on. LED panel has been bought into wide use as a means of realizing large screen display. In recent years, with the improvement of LED materials and control technology both at home and abroad, LED panel industry becomes a new high-tech industry. However, the quality testing of LED panel relies a lot on manual work in our country, which results in that the detecting speed and quality are influenced by personal opinions to a great degree.The number of points out of hand and color homogeneity are the two important indicators for quality assessment of LED panel. Therefore, this thesis mainly focused on solving the above-mentioned problems. A relatively low-cost intelligent detecting system of LED panel had been designed by bringing machine vision technology into the detecting course of the full color LED panel.First of all, a survey of LED panel development is reviewed, which emphasizes particularly on analysis of the problems existing in the production, application and detecting process of the LED panels. Then the development environment of machine vision system LabWindows/CVI is introduced, and the hardware design of true color panel detecting system directed at practical applications is proposed. And then the color homogeneity of LED panel is discussed in color theory terms, and a solution to adjust the color homogeneity of LED from the aspect of tricolor coordinate and white balance is put forward. In the end, the software of panel detection is designed based on the image segmentation, image positioning&marking and image matching technology. Simulation results prove that this system can verify the out of point accurately and adjust the color homogeneity of LED panel in a certain range.