Research Of Image Distortion Correction And Subpixel Rendering Technologies For Silicon-based OLED Display In VR

VR(Virtual Reality)is developing rapidly with the support of 5G and cloud computing technologies,and its development trend can be predicted to be energy-saving,portable,and high-definition.Therefore,the silicon-based OLED(Organic Electroluminesence Display)based on sub-pixel rendering has become an important solution for VR displays due to its characteristics of small area,low power,and high definition.However,sub-pixel rendering technology often introduces edge noise,so it is meaningful to develop a noise-reducing subpixel rendering algorithm.In addition,since the assembly error of the VR headset will introduce image distortion,it is necessary to design a low power,low resource,and high performance image distortion correction scheme for VR scenes.This paper proposes an image distortion correction scheme and a noise suppression subpixel rendering scheme according to the needs of VR scenes,and uses Verilog to complete the hardware implementation.The contribution of this paper can be summarized as the following four points:(1)In order to reduce the computation in traditional coordinate mapping algorithm,a coordinate mapping algorithm based on differential point-by-point scanning is proposed by using the feature that the mapping points of pixels in a same row are arranged equidistantly and collinearly,this algorithm can reduce the computation of coordinate mapping from 8 multiplications(divisions),6 additions(subtractions)per pixel to 0 multiplication(division),2 additions(subtractions)per pixel;(2)The Verilog design of the image distortion correction module is completed,which includes the coordinate mapping module,gray-scale reconstruction module and line buffer.This paper first uses the proposed coordinate mapping algorithm to build the coordinate mapping module,then uses the spatial locality of memory access to design a locality line buffer,which can reduce the theoretical memory overhead by 90.8%,and finally proposes a single line buffer access design by using the three-level register cache structure,and based on this design,the gray-scale reconstruction module is implemented,which reduces the memory access frequency from 2-4 times per cycle to 0.5-1 times per cycle.(3)An adaptive sub-pixel rendering algorithm based on negative compensation is proposed and implemented in hardware.The algorithm can effectively suppress edge noise according to the characteristics of the input image.The verification results on Matlab show that the proposed coordinate mapping algorithm can improve SSIM(Structural Similarity)by 37%,and the subjective correction effect approval rate can reach 97.3%.The comparative analysis on FPGA platform shows that the computing power consumption of the image distortion correction scheme in this paper is reduced by an average of 92.2%,the resource overhead is reduced by an average of 52%,and the maximum frame rate that can be supported is increased by an average of 4.2 times.By using Kodak lossless true color image set to verify the proposed sub-pixel rendering algorithm,it's concluded that the algorithm can improve SSIM by an average of 5.4%.In summary,the image distortion correction scheme proposed in this paper has the characteristics of low power consumption,few resources consumption and high performance,and the proposed sub-pixel rendering algorithm has good noise suppression performance.