Research On Multi-Channel Parallel Transmission Technology Of Real-time Video Based On ROI

With the development of video transmission technology,users have higher requirements for video picture quality.High-definition videos such as 4K provide consumers with a better user experience through finer texture depiction.High-definition video will be the development trend of next-generation video,but high-resolution will increase video clarity while also increasing the video bit rate,bringing more visual redundancy.Based on the region of interest(Region Of Interest,ROI)video coding method,the image is divided into ROI and non-ROI,and non-ROI is given a larger quantization parameter(QP)to reduce the non-ROI output bit rate,and The saved bit resources are allocated more to ROI,thereby improving the video quality after ROI encoding,and finally achieving encoding optimization in a specific target scene and improving the quality of video service.Although ROI coding technology uses a larger quantization step size for rough coding in the background area,it can effectively save bit resources and reduce the output bit rate,but due to the limitations of the network infrastructure bandwidth,a single wireless network still cannot provide users High-quality real-time video services.Fortunately,in the next-generation network,heterogeneous and converged wireless networks use different networks for parallel data transmission,which can provide users with more efficient and stable real-time video transmission services.Therefore,it is of great significance to study ROI-based real-time video multiplex parallel transmission technology under heterogeneous networks.Based on the existing multi-link transmission model,this paper proposes a new multi-link transmission architecture based on the video stream of the region of interest,which solves the problem of video based on the region of interest in heterogeneous wireless networks under the constraints of bandwidth and delay.Streaming problem.Then,according to the proposed transmission framework,a delay model and a rate control model are established respectively.Based on the model,a heterogeneous network is proposed for the adaptive rate control and bit resource allocation problems in real-time video transmission under heterogeneous networks.Next is a rate control algorithm based on the region of interest.The algorithm has been tested and verified by the actual system,which proves that the proposed rate control method can effectively improve the user’s subjective visual quality.When the delay constraint is 250 ms,the frame loss rate is less than 7%.Compared with the comparison algorithm,the algorithm in this paper The image weighted peak signal to noise ratio(Weight-Peak Signal to Noise Ratio,Weight-PSNR)is increased by up to 5.6d B.In addition,this article builds a multi-channel parallel transmission system for video streams in the region of interest based on the network simulation device Holo WAN under the Ubuntu 16.04 system.The system can detect and obtain the macroblock coordinate information of the region of interest,and on the basis of effectively estimating the bandwidth and delay of LTE and Wi Fi networks,it performs macroblock-level adaptive bit rate control to control the video bit rate more stably at the target code.Rate nearby.In terms of delay performance,the multiple parallel transmission system of the region of interest video stream is better than the single link environment.Specifically,the instantaneous delay under the ROI multi-channel parallel transmission system is reduced by up to 69 ms compared with that under the single link,and the average delay is reduced by up to 20.32 ms compared with the single link environment.