The Research Of Automatic Optical Detection Algorithm On LCD Panel Of IC Area

With the development of science and technology,the application field of liquid crystal display device is more and more widely,at the same time,the market has higher requirement for the quality of LCD panel.Whether the conductive channel of IC area has defect or not,which determines the quality of LCD panel.At present the defect detection of LCD IC area rely on manual detection in domestic,which is single.The factories observe it under the microscope and judge with the naked eye.This method is time-consuming and laborious,and due to the requirement of production line speed,it can’t do full inspection.This thesis based on the digital image processing techniques and a defect detection algorithm of IC area is designed,which can be applied to automated defect equipment.It can make up the shortcoming of manual detection.The image acquisition system is introduced in the first place.The optical structure is based on the principle of the differential interference due to imaging for the subtle conductive golden ball.At the same time,the line-scan digital camera and blue light will be used for getting a better image quality and meeting the real-time requirement of industrial production.The focusing algorithm must be used because the depth of field is less than 20 um.Four focusing algorithms of time domain and an algorithm of frequency domain are contrasted.Considering efficiency and precision,an algorithm based on brenner gradient will be used.Then the image preprocessing,the deflection angle of the image is calculated by template-matching and the ostu algorithm.A bilinear interpolation algorithm is adopted for image rotation according to the angle.At the last an algorithm of ITO segmentation based on recursive thought is proposed,it has a low computational complexity and can apply to all kinds of ITO.In the end,a defect detection algorithm to using the ratio of local mean pixel value as the basis of judgment is designed,whose error rate is less than 1% and the miss rate is less than 0.2% by testing two sample groups.It meets the detection requirements for industrial production.At the same time the OpenMP technology is used to accelerate the algorithm,which makes the time of auto-focusing algorithm shorten three times,the time of ITO segmentation algorithm shorten one time,the time of defect detection algorithm shorten three times.So the acceleration effect is obvious.The characteristic of this thesis is that an algorithm of ITO segmentation and a defect detection algorithm are proposed,which have high detection accuracy.Finally the algorithm is accelerated by OpenMP technology and the effect is obvious.