| With the rapid development of society,high-definition(HD)video plays an important role in many fields,such as broadcasting and television,security monitoring,medical and health care,intelligent transportation,aviation and aerospace.The acquisition and transmission of mass HD video are inseparable from video coding technology.Based on the relevant projects of our research group,this paper studies the lossless compression technology in the imaging module and transcoding technology in transmission system,whitch are according to the technical requirements of the acquisition end and transmission end in the HD video application.The main work of this paper is as follows:1.This paper proposes a low complexity compression scheme for Bayer images generated at the imaging end of digital camera,which achieves compression performance close to the mainstream standard under the premise of a small amount of resources,and implemented it based on FPGA.2.Aiming at the demand of efficient transmission of a large amount of video data,a real-time multi-channel HD video transcoding system was designed,and a series of tests were completed,then it was applied in enterprises.With the increasing of video resolution,the size of Bayer images captured by the front end of HD digital camera becomes larger and larger,which makes the storage resource required for image processing modules soar.Lossless compression of Bayer images can effectively reduce the cost of imaging device and avoid error porpagation in the subsequent processing.Some existing image coding technologies are studied in detail in this paper.According to the characteristics of Bayer image,a low complexity Bayer image lossless coding method is designed and implemented based on FPGA.First,we divide Bayer images into several sub-image blocks,then four cross-pixel prediction models are used to eliminate the spatial redundancy of the image.After that an improved adaptive Huffman coding method is proposed to encode the predicted residuals and output the code stream.Finally the coding of Bayer image is completed block by block.Experimental results show that the compression performance of our method is close to that of the exiting image coding standards.Bayer images continuously collected by the front end of the camera go through a series of processing,and the video data formed is transmitted and stored after lossy coding.In practical applications,the massive multi-channel HD video caused great pressure on network resources,in addition to this,models and standards of camera devices are different,which challenges the stability of the system.Transcoding the video according to requirements in the transmission node can flexibly adjust the encoding attributes of each video,so that network resources can be used more appropriately.Based on such application requirements,this paper carries out research on video coding technology in transmission nodes.We first studies key technologies of the existing video coding standards such as H.264/AVC,H.265/HEVC and H.266/VVC.Then based on the practical requirement of enterprise application,we designed a multi-channel video real-time transcoding system.By taking the requirement of low complexity and high reliability into account,we constructed the development environment and onboard running environment based on HI3531D media processor.The architecture and modules are designed according to specific indicators,and a set of hardware and software system are developed to meet the needs of enterprises.The main functions of our transcoding system include:multi-channel HD video streaming,real-time decoding and encoding,bit stream push,web front-end service,ONVIF specification and video storage and playback.The system has been applied in the cooperation enterprises of the research group,and won the award of the 9th China Innovation and Entrepreneurship Competition(Shandong). |
PDF Full Text Request