Study On The Delay Optimization Of A Video Transmission System Based On Intelligent Terminals

With the development of network communication technology and the popularity of portable intelligent mobile devices, the application of real-time video transmission system based on intelligent terminals becomes increasingly common.However, due to the limitations of hardware devices and implementation approaches,the real-time video transmission system based on intelligent terminals often has the problem of large delay and video playback instability, which makes the video transmission performance can’t meet the design requirements, and the user experience is poor. Based on the above problems, this paper proposes the optimization scheme through the detailed study of the system’s acquisition terminal and server. This paper implements and tests the system, verifys the effectiveness of the optimization scheme.Firstly, through the research on the acquisition module of the system’s acquisition terminal, an initialization algorithm is proposed to match the acquisition rate and coding rate to avoid encoding delay. Through the research on the the coding module of the system’s acquisition terminal, this paper recompiles the FFmpeg library and only retains the H.264 encoding function to improve the coding efficiency and reduce delay. At the same time, the code rate control method, GOP and other coding parameters are optimized to achieve the optimization of the coding module.Then, through the research on the data buffer of the server, this paper proposes a real-time processing method based on playback connection to reduce the delay.Through the research on the process of RTP packaging in the live555, this paper proposes a RTP packing algorithm which is based on the size and type of NAL units,improves the packing efficiency and reduces delay. This paper proposes an adaptive retransmission mechanism to avoid the sudden interruption during real-time playback. The server sends feedback to the acquisition terminal, and the acquisition terminal dynamically adjusts the transmission rate based on the feedback, which improves the stability of video transmission.Finally, this paper designs and implements the above schemes. The real-time video transmission test result shows that the average time delay reduces to less than1 s and can meet the real-time requirements. The stability of video transmission is also significantly improved.