复杂无线信道环境下渐进图像的信源信道联合编码

恶劣无线信道上丢包和随机误码共存,使得渐进图像的鲁棒高效传输异常困难。不同于以往基于乘积码和Reed-Solomon(RS)码的编码保护方案,该文提出了基于Raptor码的信源信道联合编码方案。该方案采用Raptor码对脆弱的压缩码流提供不等重保护来抵抗信道的丢包和误码。为了最小化端对端传输失真,利用动态规划算法,求得信源压缩和信道不等重保护的最优码率分配方案。仿真表明:相比于以往的传输保护方案,该方案使得端对端的图像重建的峰值信噪比提高了约2～5dB;同时搜索最优码率分配方案的复杂度和信道编码复杂度有了明显降低;解码的端对端的延迟也大大减小。

Joint source-channel coding of progressive images for complicated wireless channel environments

The coexistence of packet erasure and random bit errors make robust and efficient transmission progressive images on wireless fading channels quite difficult.Previous methods have used Reed-Solomon(RS) codes or product codes to improve transmissions.This paper presents a joint source-channel coding method based on Raptor codes to protect vulnerable compressed streams from both packet erasure and random bit errors.End to end distortion is minimized by a dynamic programming algorithm to obtain the optimal rate allocation between source coding and unequal error protected channel coding.Extensive simulations show that the method provides significant peak signal to noise ratio(PSNR) gains(2~5 dB) over previous protection methods,that the complexity of channel encoding and searching for the optimal rate allocation scheme can both be greatly reduced,and that the decoding delay is reduced.