图像传输信源信道联合解码算法  

Joint source-channel decoding scheme for image transmission

作　　者：谢旭东[1] 裴玉奎[1] 陆建华[1] 

机构地区：[1]清华大学电子工程系,北京100084

出　　处：《清华大学学报（自然科学版）》2008年第10期1602-1605,共4页Journal of Tsinghua University(Science and Technology)

基　　金：国家"九七三"重点基础研究基金项目(2007CB310600);国家自然科学基金资助项目(60532070;60525107);中国博士后科学基金资助项目(20070410060)

摘　　要：为提高遥感图像传输的可靠性,结合抗差错算术码和低密度奇偶校验码,提出了一种基于遥感图像压缩算法的联合信源信道解码算法。在该算法中,低密度奇偶校验码提供初步的解码结果,以帮助算术码解码器进行高效地序列解码;算术码修正后的比特信息被反馈至低密度奇偶校验码进行迭代解码,直至置信度解码算法收敛或者抗差错算术码算法中的堆栈为空(此情况下即宣告解码失败)。仿真结果表明,该方案同分离解码算法相比,能够有效地降低传输噪声的影响,重建图像的峰值信噪比最大能够取得15 dB以上的增益。An efficient joint source-channel coding （JSCC） scheme was developed to improve the reliability of remote sensing image transmissions. The scheme combines error resilient arithmetic codes （ERAC） and low-density parity-check （LDPC） codes for remote sensing image compression. The initial decoding results from the LDPC decoder are sent to the ERAC decoder for an efficiently sequential decoding. Then, the corrected information provided by the ERAC decoder is sent back to the LDPC decoder for iterative decoding until the belief propagation （BP） algorithm converges or the ERAC stack is empty, in which case, the decoding fails. Test results show that the approach more effectively reduces the effect of noise than the separated decoding technique, with more than 15 dB increase in the peak signal-to-noise ratio （PSNR） of the reconstructed image.

关 键 词：图像编码 联合信源信道编码 抗差错算术码 低密度奇偶校验码 

分 类 号：TN919.81[电子电信—通信与信息系统]