Research On Video Precoding Technology Based On H.264/SVC

In recent years,the demand for mobile short video services has grown exponentially,and there is an urgent need to increase the service rate of mobile video services to meet the viewing quality requirements of users.However,due to the scarcity of spectrum resources,the wireless communication resources are limited,so that the increase in service rate cannot keep up with the user’s demand for video services.Therefore,how to match video coding with wireless communication link transmission has become a hot topic that has been widely concerned.On the one hand,massive digital video needs to be effectively compressed in storage,otherwise it will take up huge storage resources.On the other hand,in order to obtain high-quality video,it must be effectively and reliably transmitted under the condition of limited network resources.This requires joint processing of video compression and video transmission to match the video with network resources.Video compression is not a new problem.It has developed many mature coding standards after years of development.Among them,the H.264 video coding standard adopts a hybrid coding framework of predictive transformation,which not only achieves good compression efficiency,but also has certain network affinity.And its extended version H.264/SVC scalable video coding protocol is tailored for video transmission under heterogeneous networks.However,these current video compression standards and frameworks do not systematically study the dynamic changes of wireless network resources,which makes the encoded video mismatch with time-varying network resources,resulting in poor received video quality.In order to enable users to get a better video viewing experience,research on video compression and video transmission technology is urgent.In view of the above scenarios,this thesis first studied the characteristics of the scalable coding structure of time,space and quality in the H.264/SVC coding protocol,and found that scalable video coding has a clear "cliff effect".For this problem,the relationship between the number of video coding layers and the quality of video reception and the robustness of the temporal scalability,spatial scalability,and MGS structure(medium grain scalable,MGS)in the quality scalability are analyzed in depth.A video precoding technology based on SVC is proposed.The specific research content is as follows:Firstly,according to the demand of video service,the relationship between heterogeneous cellular network and video transmission is analyzed.Due to the complex networking of heterogeneous networks,diverse network equipment,and large differences in storage and computing capabilities,video encoding and transmission are severely limited.Therefore,how to effectively encode and reliably transmit has become a hot issue of video processing technology.Due to fluctuating bandwidth,unreliable channels and diversification of equipment,the encoded video stream should have the ability to adapt to dynamic changes in bandwidth,and be able to correct errors and detect errors,and adapt to different user equipment features.Secondly,since the SVC technology can overcome the disadvantages of the fixed and single bit rate of the traditional video coding stream,it has certain adaptability to heterogeneous cellular networks.This thesis is based on SVC under the H.264/AVC coding standard,and realizes three common coding methods of time,space and quality through the SVC reference code JSVM.The advantages and disadvantages of SVC coding structure are analyzed.The study found that although SVC changed the previous single-layer coding method,it adopted a multi-layer coding structure to better adapt to the heterogeneous network environment and the diversity of user equipment.However,SVC still has an obvious "cliff effect" and cannot adapt to the dynamic changes of network resources.Especially when network resources are scarce,each layer of video data still needs to be transmitted,which increases the burden on the network.Thirdly,In view of the shortcomings of SVC,this thesis starts with the structure of the three scalable coding methods of SVC,and analyzes the robustness of different scalable coding structures.A layered-oriented mobile video precoding algorithm is proposed.Its main idea is:according to the feedback CSI(Channel State Information,CSI),precoding is first performed in the encoded bitstream,and the number of enhancement layers of video transmission is dynamically adjusted.Such preprocessing can not only effectively save network resources,but also be able to Get better video quality.Further,on the basis of optimizing the number of enhancement layers,in order to minimize the problems caused by the "cliff effect",a more refined precoding method for video data blocks is proposed.Its basic idea is:to discard some lower priority data blocks to adapt to the network conditions,and retain higher priority data blocks to ensure the quality of decoded video.Compared with the layer-oriented precoding method,it can dynamically adjust the number of NLAU for transmitting video streams.Such preprocessing not only inherits all the advantages of layered preprocessing,but also makes full use of network resources.Finally,Corresponding experiments are given to verify and demonstrate the above two algorithms.Experimental results show that the two precoding methods proposed in this thesis can effectively improve the video reception quality at the decoding end.