Research And Design Of Streaming Media Video Compression Algorithm Based On Interframe Correlation

In recent years,with the widespread popularity of the Internet and communication equipment,the network has become one of the important channels for people to obtain all kinds of information,therefore,a variety of multimedia applications have been generated.In this process,the network bandwidth affects the speed of information transmission,and the speed of information transmission is an important factor affecting the quality of people's experience,so streaming media technology emerged as the time require.However,affected by the huge amount of data and high requirement of real-time data,fast growth of data,high frequency query factors,how to effectively compress the video streaming media,so as to realize the storage and transmission of information is a problem needing to solve urgently.The traditional Shannon sampling theorem uses the signal acquisition method that sample first and compress later,which results in a waste of resources.Compressive sensing(CS)theory compresses the information while this is sampled,greatly reducing the complexity of encoding and the waste of resources,so it has been widely used in video encoding field.This video coding scheme based on compressed sensing theory is called Compressive Video Sensing(CVS)scheme.The traditional CVS algorithm usually uses the same measurement matrix for each frame image to measure and reconstruct.Aiming at the obvious interframe correlation of video signals,a CVS scheme based on interframe correlation is proposed in this paper.The main research contents and work are as follows:1.A block video coding scheme based on bidirectional prediction and variable sampling rates is proposed.Firstly,the reference frame is sampled through fixed and high sampling rate,and then the non reference frame is divided into nonoverlapping blocks and classified according to the degree of change of the image blocks,and different sampling rates are allocated to different types of image blocks.In the process of image block classification,first prediction is done by bidirectional prediction,then the intra frame correlation is used to correct the initial classification results.The experimental results show that this method can improve the classification accuracy of the image block effectively,and make the classification result closer to the actual situation.Thus,this method can reduce the sampling rate of frames to a certain extent and improve the quality of video reconstruction effectively.2.A multi hypothesis prediction video decoding scheme based on variable sampling rate is proposed.Firstly,the reference frame can be directly reconstructed because it measured with fixed high sampling rate.For the non reference frame,firstly it is necessary to preprocess two types of image blocks with low sampling rate.Secondly the prediction frame is predicted by multi hypothesis.Thirdly a more accurate prediction frame can be obtained through bidirectional motion estimation.Finally,the error between the predicted frame and the original frame is corrected by residual reconstruction,and a high quality reconstructed frame is obtained.The experimental results show that this video reconstruction scheme can reconstruct and obtain high quality video frames at lower sampling rates.