Research And Application Of Fast Vedio Encoder And Control Algorithm

Along with the widespread popularity of Internet video-on-demand applications, a new generation of highly efficient video coding standard(High Efficiency Video Coding, HEVC) launched, high-definition video has become mainstream video development trend. However, the speed of network bandwidth enhance so slow that far behind with streaming HD video application requirements. How to quickly complete the video encoding transmission in limited bandwidth, and to ensure maximum visual, is both a hot and difficult research topic for a long time. In this thesis, our study focus the key to solve the problem of video coding transmission under low-bandwidth environment, mainly from these three aspects: cutting the encoder, bit rate control and video image distribution optimization under low-bandwidth.(1) Realize the rapid HEVC video encoder. We got the rapid HEVC video encoder through the analysis of HEVC coding structure, coding complexity, experimental testing different coding techniques for video quality improvement and the contribution of video encoding time consumption, then cut off part of the function. The results of experiment show that: the rapid HEVC video encoder save the encoding time by 99.28% on average, speed up more than 150.19 times with an averge 75.99% increase in BD-rate, compared to the standard HEVC video encoder.(2) Propose a bit rate control algorithm under low-bandwidth. Bit rate control algorithm under low-bandwidth includes scene changing control and adaptive variable frame rate control. Scene changing control detect the scene changing by sample. Adaptive variable frame rate control adjust the frame rate by combined QP value with a buffer occupancy as criteria. The experimental results show that the proposed low bandwidth bit rate control algorithm can quickly switch accurate detection scene changing, can make the buffer of time at a high occupancy rate with stable and excellent output video quality.(3) Propose a video distribution optimization algorithm under low-bandwidth. Research distribution optimization algorithm is based on the analysis of HEVC standard reference rules and improve to eliminate dependencies between the GOP, then propose a mathematical model describing the inter-frame structure with a binary reference dependencies. Finally, the video image distribution problems of different coding modes are converted to 0-1 knapsack problem. Experimental results show that the new video image distribution optimization algorithm revenue more than hierarchical image distribution under all bandwidths.Based on the research of the above three aspects, this paper provides a complete set of solutions for lower bandwidth video encoding, including the implementation of rapid video encoder, propose a bit rate control algorithm and the video image distribution optimization algorithm. When coding and communication with the network bandwidth resources is not sufficient, where having a high real-time demand, the paper’s research results will have significant practical application value.