OLED Mobile Phone Screen Brightness Non-uniformity Extraction Algorithm

In recent years,as a new display medium,the OLED display has been widely used in life due to its environmental protection,energy saving and durability.With the increasing demand for OLED displays in the market,while improving the corresponding manufacturing process,the pre-factory inspection and brightness correction links are also the technical keys to improve the yield.At present,the use of machine vision technology to build a brightness detection platform for compensation correction is an effective means to improve the efficiency and quality of display screen inspection and correction.This thesis mainly studies the brightness extraction and processing algorithm of the pixels of the OLED mobile phone screen based on machine vision(hereinafter referred to as the light point to distinguish from the image pixels),and the final calculation results are used by the compensation circuit to correct the brightness non-uniformity of the display screen.In view of the imaging characteristics of OLED mobile phone screen light points,this thesis proposes a detection method combining gray information and gradient changes and a simple and effective incident light integration method.The extracted data can reflect the actual brightness of the OLED light point and can adapt to different resolutions,multiple The location and brightness calculation of mobile phone screen lights based on RGB light points.Aiming at the difficulty of arranging and locating the lamp points due to the different sizes,contours and shapes of mobile phone screens and possible defects such as chipping during the cutting process,an adaptive search lamp strategy based on the area growth method was proposed.Coordinate prediction and brightness filling are performed on the located abnormal light point.This algorithm can solve the problem that it is difficult to search and mark the lamp points because of the different contours of the mobile phone screen.In order to improve the uniformity of the OLED display screen as much as possible under the limitation of the limited bit width of the compensation circuit,the spatial domain high-pass filter is used to suppress the low frequency and retain the high-frequency and high-frequency non-uniformity;for the brightness data of the curved area on both sides of the curved screen Darkness affects the compensation effect.The surface area data is adjusted to approximate gray level with the plane area.The piecewise linear filter window is used to solve the problem of surface boundary compensation distortion.In the research of brightness data standardization,combined with the imaging principle of the CCD camera,the process from the photon emitted from the mobile phone screen to the final imaging is analyzed,and different conversion coefficients are set for different RGB channels to solve the color cast phenomenon on the mobile phone screen after compensation problem.In terms of improving the real-time performance of the algorithm,the CUDA programming technology is used to accelerate the process of positioning and brightness integration that form a calculation bottleneck.The finally developed algorithm runs on the machine,and the compensation effect and real-time calculation speed achieved by it have been approved by the OLED screen manufacturer and will be applied in the production line soon.The research results show that for mobile phones with different styles and multiple RGB arrangements,the brightness non-uniformity extraction software can accurately locate the light point and extract the brightness,and realize the parallelization of the algorithm through the combination with CUDA.Efficient features.