The Research Of Detection Device Of Lcd Screen Character Defects Based On Machine Vision

The LCD screen is more widely used in people’s daily life. Whether the display of LCD character is correct or not has become an essential testing process for some products in the production process. The traditional detection methods mainly depend on manual testing methods. Manual testing methods are in low efficiency and have strong subjective. So the article started a research for the LCD screen character defects detection based on machine vision. The overall design of the detection system and the corresponding defect detection algorithm were discussed. The automatic detection method of LCD character defects achieved. The article did the following works:Based on machine vision, the overall scheme of detecting device was designed according to the needs for the LCD screen character defect detection. Through analysis and selection of the image sensor, lens and the light source, a set was designed which mainly including the image acquisition module, the illumination module and the computer. The mechanical drawing of the device was done and helped complete the processing of the detection system’s mechanical device.The extraction algorithm for the required template in template matching was studied. The algorithm included filtering and thresholding. Filtering was mainly to suppress image noise and highlight the parts to be extracted. The filtered image was then segmented by image threshold binarization. The required template image was gained.The algorithm for the LCD character defect detection based on machine vision was studied. The core of this algorithm was the image matching algorithm based on the shape template. An improved method of similarity measurement was proposed to overcome the light effect on the matching results. By image matching, the template image instance was found in the images which to be tested. After determining the character position, the characters were extracted and then to obtain the number of pixels of them. Finally, the results were given out by comparing with the set value.The software detection system was designed in Microsoft Visual Studio 2005 integrated development environment and image processing software Halcon, whichrealized the automatic detection function. With the LCD screen samples provided by the manufacturer, experiments were carried out respectively using this algorithm,based on SIFT matching algorithm and artificial method three ways. The experimental results showed that this algorithm was able to accurately detect the LCD screen characters’ defects and did the judgment. Compared with the other two methods, the accuracy and efficiency of detection were both improved. It can replace the existing manual detection method.