Research On High-Precision Substrate Angle Detection And Correction Algorithm For Inkjet Printing

Organic light-emitting devices(OLED)have become the mainstream display devices in recent years because of their low energy consumption,high brightness,high contrast,rapid picture response,and flexible processing,etc.They are widely used on display devices such as mobile phones,TVs,tablets,smart wearables,etc.In the manufacture of OLED key functional layers,compared with the traditional evaporation process,inkjet printing technology is considered to be one of the most potential film-forming technologies for OLED panels because of its characteristics such as no need to use masks to achieve patterning,high material utilization,simple process and low cost,and it has also become one of the key technologies of "Made in China 2025".In the inkjet printing process,how to perform real-time,high-efficiency,highprecision multi-target positioning and angle detection on the high-resolution pixel pit substrate has become an urgent problem to be solved.Aiming at the problem of high-precision substrate visual positioning and inspection,this paper proposes a multi-target edge template matching algorithm based on Gaussian pyramid.The main research contents are as follows:(1)Based on the overall accuracy requirements of the inkjet printing system,the corresponding visual inspection accuracy index was formulated,and on this basis,the visual inspection system is designed and built.According to the characteristics of the inkjet printing process,the algorithm of substrate location detection and correction for pixel pits in inkjet printing is designed and implemented.On this basis,a visual detection software system for inkjet printing is developed,which realizes the functions of substrate detection,angle correction and equipment control,and is applied in the actual ink-jet printing experimental equipment.(2)To improve the Canny edge detection algorithm,an adaptive Canny edge detection algorithm based on the maximum inter-class variance is proposed,which improves the accuracy of edge detection and reduces the number of edge points.In the pyramid of multi-scale and multi angle images,the algorithm can select high and low thresholds adaptively according to different image features,which improves the stability and adaptability of edge detection.This method is used to generate an edge template pyramid,and this pyramid is used for edge template matching,which greatly reduces the number of matching calculation points,thereby reducing the amount of calculation in the entire matching process.(3)A multi-target edge template matching algorithm based on Gaussian pyramid is proposed,which is used for pixel pit location and substrate angle detection.According to the change characteristics of the normalized vector dot product with the angle of the vector,it is applied to the edge gradient direction detection of the matching image,and then the matching angle detection is realized.For matching in pyramid,a variety of acceleration optimization strategies are proposed to improve the calculation efficiency and positioning accuracy of the algorithm.Finally,a substrate correction algorithm based on step optimization and a straightline fitting angle verification method for verifying the substrate correction results are proposed.(4)The software and hardware platform of the inkjet printing visual inspection system is designed and built,and the accuracy,efficiency and feasibility of the substrate angle detection and positioning algorithms based on template matching proposed in this paper are verified through experiments.The algorithm of this paper is integrated into the inkjet printing software system for actual printing test.The printing experiment results show that the overall algorithms of this paper have good real-time and robustness,and can be applied to the inkjet printing process correctly and appropriately.In summary,aiming at the requirements and problems of high-precision and highefficiency detection of high-resolution pixel pit substrates in inkjet printing systems,this paper proposes a multi-target edge template matching algorithm based on Gaussian pyramid.By improving the edge detection algorithm and the edge template matching algorithm,the template matching is used to detect the angle of the object,and the Gaussian pyramid algorithm is used to accelerate the algorithm,which realizes the high-precision and high-efficiency visual positioning detection in the ink-jet printing system process.