Research On Video Scaling System Based On FPGA

With the rapid development of 5G and artificial intelligence,video transmission has also changed from traditional standard definition and high definition to ultra high definition.As the resolution of video becomes higher and higher,the amount of data it transmits also increases dramatically,which requires a very short transmission response time.So in the field of digital video,video data synchronization is a key problem to be solved urgently.With the popularity of ultra high definition video sources,the resolution of transmitted video is getting richer and richer.In order to meet the diverse demands for video display effects in scenes such as intelligent oil fields,rail transit and large exhibition centers,a hardware transmission system for real-time video scaling is needed.Aiming at the above problems,this paper proposes a research method of video zooming system based on field programmable gate array(FPGA).Firstly,different interpolation algorithms are simulated and compared by MATLAB,and the characteristic curves of these interpolation algorithms in spatial and frequency domain are analyzed one by one.At the same time,after weighing the hardware complexity of the algorithm,the hardware resources of the chip and the effect of image scaling,we finally choose to implement the double cube interpolation algorithm on FPGA.Compared with PC,FPGA can be used to obtain higher processing speed,and its pipeline technology is used to output the pixels after interpolation operation in turn,so as to fully realize the parallel processing mode.The system mainly consists of five modules: video image adaptive acquisition module,zooming parameter configuration module,zooming module,video data cache control module and output control module.The key of its design is to control the coordination of each hardware module.The system makes full use of the pipeline design of FPGA.In addition to the main function of video image scaling,the system also realizes the adaptive collection of video source,the custom configuration of video image scaling resolution,and the reconfiguration of output video,etc.The key of system design is the coordination between hardware modules.The system finally realizes custom real-time zooming and display at 1080 P and below resolution.