| Education sector is always a major concern of people.With the development of our economy and growth of living standard,education investment in family consumption increases greatly.Education related industry develop fast.Information technology such as internet develops really fast in recent years.Online education combines information technology and internet business mode,which generates lots of online education forms,including open class,video on demand,live video and so on.These online education products depends on digital video transmission system.A good digital video transmission system is vital to online education quality.A digital video transmission includes video codec and transmission modules.Video coding mainly focus on digital video signal compress coding to save internet bandwidth.HEVC is the new generation video coding standard.Compared with former video coding standards,it introduces lots of new video compression technologies.It has outstanding compression rate.It also has great computation complexity because of its complex encoding process.It raise a big challenge to live video applications.Online education involves multiple peer video communication and frequent interaction,which needs efficient video/audio streaming transmission and complex signal control.This thesis did a reach on HEVC high complexity and live video transmission of online education.Motion estimation is a vital module of encoder,which compensates a big part in the whole complexity.Sub-pixel motion estimation compensates a big part in motion estimation too.This thesis did a reach on sub-pixel motion estimation optimization algorithm.Compared with original global search method,we use algorithm fitting to estimation the sub-pixel,which can effectively reduce search frequency and also maintains good compression rate.This thesis also participated in online education video transmission system.For online education scenarios,this thesis focus on signaling transmission and video transmission.We use websocket protocol to achieve signaling transmission and RTP/RTCP to achieve video stream transmission.In order to impove online education transmission quality,we propused several system level optimization methods,including signaling sequence refratering,TURN server clustering,recording system optimization based on region and so on.This video transmission has been online in an online education agent,which provides service to lots of users.This thesis introduce HEVC key technology and analyze HEVC sub-pixel motion estimation.Then it introduces three optimization method details and experiment results.Through the analysis of simulation results and comparison of standard sequence and online education sequence,under the precondition of the decrease of PSNR,we get the optimization algorithm with faster rate.Next,it introduce the online live video transmission system framework and focus on design and implement of signaling and video stream transmission.Then it introduces the system level optimization methods including signaling sequence refratering and so on.Through these optimization methods,the problems actually encountered by the online education system are solved.The performance and user experience are obviously improved,and the online education system can be put into practical use to achieve better results. |
PDF Full Text Request