A Study Of Video Active Transmission Method Based On Network Environment Perception

With the rapid development of the Internet and the gradual deepening of smart cities and intelligent transportation,the network traffic generated by video services has become the most important data contributor to the total traffic of the Internet,and video transmission technology has become an indispensable part of the information society.However,factors such as massive video data,time-varying networks,and video data redundancy all bring great challenges to end-to-end video transmission,which are summarized as: 1.Large amount of video data;2.The available bandwidth is low;3.Large bandwidth fluctuations.These three challenges restrict the performance of the video transmission system.To improve the performance of the transmission system and reduce the network transmission delay,this paper carried out the following research work:First,it addresses the slow data transfer speed due to low available bandwidth.In these studies,an adaptive video stream transport(BAAT)model based on available bandwidth awareness is proposed.The model is deployed at the endpoint,and the available bandwidth awareness model is designed to perceive the available bandwidth in real time when the bandwidth resource is high,and the available bandwidth is low.Minimizing transmission delay as the optimization objective function,and modeling video quality as a constraint to achieve feature-based compression of video at the sending end,and adaptive adjustment of video stream compression ratio transmission according to the change of available bandwidth,to minimize transmission delay.The experimental results show that the average delay of the BAAT method is reduced by 60% to 90% compared with the bitrate adaptive method and the cache adaptive method.Second,it is difficult for the receiver to receive video images in time due to large amounts of video data and large bandwidth fluctuations.This study adds the consideration of semantic feature flow for machine vision and proposes a multi-data stream active transmission model based on network environment perception.By sensing the current bandwidth size and bandwidth fluctuation amplitude,the model divides the harshness of the network environment,transmits different data flows in different harshness of the network environment,and realizes that the receiving end can receive data in different network environments and make timely decisions,to make full use of bandwidth resources and reduce network transmission delay.The research site is a unique end-to-end video transmission method with strong innovation.The experimental results also show that the multi-stream active transmission method can reduce the delay time by more than 30%compared with the traditional method of transmitting only single video streams and dual video streams.Finally,an end-to-end multi-stream active transmission experimental platform is built,and the model is implanted,which mainly uses intelligent transportation as an application scenario for experimentation.The architecture and design components of the platform are introduced,and the end-to-end transmission effect in different environments is visualized.The transmission results of the experimental platform verify the feasibility of the video active transmission model.