Research On Transmission Distortion Estimation Over Wireless Networks Based On Distributed Video Coding

Distributed Video Coding(DVC)is a new video coding scheme.Unlike the traditional video codec(such as MPEG-X and H.26X),DVC transfers the computational complexity from the encoder end to the decoder end,and has the characteristics of simple encoding and decoding complexity.This feature makes it suitable for mobile video phones,wireless video surveillance and wireless video sensor networks and other applications.In the wireless video communication system based on distributed video coding,transmission distortions caused by routing error,transmission delay,and information error,will eventually lead to the degradation of video reconstruction quality in the decoding end.Estimating accurately the video transmission distortion is the key to adjust the parameters of the system to control the quality of video transmission and optimize the overall system distortion.The existing research on wireless transmission of DVC mainly focuses on error control measures,and lacks the analysis and estimation of the distortion of wireless transmission based on DVC.Based on the research of wireless video transmission application based on DVC,this dissertation focuses on building a wireless transmission simulation platform.And then a wireless transmission distortion estimation model is established.The main work are as follows.(1)Firstly,build a wireless video transmission simulation system based on DVC.The system consists of DVC coding,video packet,wireless channel,packet loss detection,DVC decoding and error concealment module.The system divides the video stream into K frame and WZ frame.considering synthetically two kinds of distortions:K frame loss without packet loss and WZ frame packet loss,K frame and WZ frame loss at the same time;Considering the packet interleaving technique.This dissertation assumes that system packet loss is random.Experimental results show that if the rate of packet loss increases,the quality of the reconstructed video at the receiving end gradually decreases.At the same packet loss rate,the reconstructed video quality of a sequence with a high degree of motion is poorer than the quality of a reconstructed video with a slow motion.These simulation experiments can better verify the actual scenario of packet loss.The experimental results are in good agreement with the actual situation of wireless transmission,which can simulate the actual wireless transmission environment well.(2)Based on the simulation system of wireless transmission distortion in DVC,this paper proposes a non-reference(NR)transmission distortion estimation algorithm,which adopts the idea of recursion and extracts K Frame and WZ frame reconstruction information to estimate the transmission process generated transmission distortion.The algorithm estimates the transmission distortion under the condition of no packet loss of K frame,packet loss of WZ frame and packet loss of K frame and WZ frame at the same time.The experimental results show that the absolute error between the estimated value and the measured value is less than 1 dB and the relative error is less than 5%for K frame without loss and WZ frame loss.Among them,the least absolute error is 0.0069 dB,and the least relative error is 0.01%.For the simultaneous loss of K and WZ frames,the least absolute error of the proposed algorithm and the measured value are controlled within 1.8dB and the relative error is within 5%.Among them,the least absolute error reaches 0.0465dB and the least relative reaches 0.14%.The estimation algorithm proposed in this paper can effectively estimate the transmission distortion caused by packet loss during video transmission.