Research Of 1080P Low Quality Real-time Video Enhancement System Based On FPGA

At present,mobile Internet and large-scale data are developing rapidly.Video images as an important part of network information are widely used in various fields.The acquisition of video image is easily interfered by the external environment.It will affect the subsequent image processing and computer vision processing.Low-quality image enhancement research has become one of the hot spots.As the development of semiconductor technology and the arrival of large data age,people put forward the image enhancement real-time,high-definition,transmission performance and other requirements.It's an important research topic to figure out how to meet the above requirements,without loss of image enhancement effect.Firstly,this paper expounds the significance of research on low-quality image enhancement,and introduces the research status home and abroad within this field.The overall scheme design of the low quality video enhancement system is proposed,and the video processing framework and the hardware configuration necessary to meet the system requirements are determined.Secondly,the enhancement algorithms based on Retinex theory and color estimation model are analyzed in detail through theoretical analysis and concrete experiments.An improved IMSR algorithm is proposed for the deficiency of original IMSR,Retinex theory and color estimation model.In order to ensure the enhanced effect of the premise of reducing the processing time,the computational complexity of each part of the algorithm is analyzed and the computation module of the integrated surround image with high computational complexity is improved and optimized.Thirdly,the analysis of the influence of low quality image enhancement noise is taken and the principle of bilateral filtering is introduced.Improve the filter coefficient calculation method in the algorithm in order to further enhance the efficiency of the algorithm.Subsequently,the logic circuit of improved IMSR algorithm is implemented on the FPGA using Vivado HLS high-level synthesis tool,and the transplant of the algorithm is realized.The optimized implementation of the improved bilateral filtering algorithm is implemented on FPGA using Vivado HLS high-level synthesis tool,which can meet the requirement of real-time processing.The depth optimization of memory access,data transmission and image processing pipeline operation is introduced,and the design of additional software functions,such as enhanced video enhancement level adjustment and enhanced switch,is designed with MicroBlaze.At last,the improved IMSR enhancement algorithm based on FPGA is verified by RTL simulation,and the comparison chart of the image enhancement before and after image enhancement is given.According to the test results,the low-quality video enhancement system achieve the expected 1080 P low-quality video image enhancement function,processing speed,good effect,system stability.