Research And Verification Of Video Noise Reduction Algorithm Based On Spatio-temporal

With the gradual development of social science and technology,the public has higher and higher requirements for the quality of video image information,and the image signal processing technology(Image Signal Processor,ISP)has become more and more important.Video noise reduction technology is an indispensable part of ISP digital image processing,which can not only improve the user’s viewing experience,but also affect the pressure of subsequent video image processing technology.Therefore,video image noise reduction technology has a wide range of practical application value,and major ISP chip design companies have begun to focus on researching more effective video noise reduction algorithms.Starting from the actual needs of the project,this thesis studies the principles,advantages and disadvantages of different noise reduction algorithms in detail,and studies a spatio-temporal noise reduction algorithm(3D Noise Reduce,3DNR)to denoise video images.In the research process,the performance of the algorithm is improved according to the noise reduction effect of the actual algorithm,including the noise stabilization(Variance Stabilization Transformation,VST)processing and the addition of edge detection combined with Non-Local Means Filtering algorithm(NonLocal Means,NLM)to adjust together The weight of the pixel points,and the innovative proposed bilateral filtering algorithm with direction,adding the detection of the moving edge of the front and rear frames to adjust the weight of the edge part,which has a better edge preservation effect for video image noise reduction.Based on the noise reduction algorithm in the spatio-temporal,the hardware structure is designed,and a verification platform based on Universal Verification Methodology(UVM)is built for the algorithm and hardware design,and the hardware code and algorithm are logically verified using the platform.and DEBUG,modified for simulation errors.Then,the simulation verification results are collected and the coverage results are displayed.Finally,based on the coverage results,the entire verification work is summarized and analyzed,so that the coverage meets the project verification requirements and the verification work is completed.Finally,the algorithm is used to denoise the video,and the noise can be significantly removed by comparing the video images before and after denoising.The research results of this paper will help to promote the research of video image related noise reduction algorithms,lay a good foundation for the research on UVM verification platform construction,and also have certain practical application value in ISP digital image processing technology.