基于压缩感知的SAR图像鲁棒编码传输

发布时间：2018-06-07 10:39

本文选题：压缩感知 + 合成孔径雷达　；参考：《中国图象图形学报》2014年11期

【摘要】：目的尽管传统的联合信源信道编码方案可以获得高效的压缩性能,但当信道恶化超过信道编码的纠错能力时会导致解码端重构性能的急剧下降;为此利用压缩感知的民主性提出一种鲁棒的SAR图像编码传输方案,且采用了一系列方法提高该方案的率失真性能。方法考虑到SAR图像丰富的边缘信息,采用具有更强方向表示能力的方向提升小波变换(DLWT)对SAR图像进行稀疏表示,且为消除压缩感知中恢复非稀疏信号时存在的混叠效应,采用了稀疏滤波方法保证大系数的精确恢复,在解码端采用了高效的Bayesian重建算法获得图像的高性能重建。结果在同等码率下,与传统的联合信源信道编码方案CCSDS-RS相比,本文方案可以实现更加鲁棒的编码传输,当丢包率达到0.05时,本文方案DSFB-CS获得的重建性能明显要高于CCSDS-RS;与基于Bayesian重建算法TSW-CS的传统方案相比,本文方案可提高峰值信噪比(PSNR)3.9 d B。结论本文方案DSFB-CS实现了SAR图像的鲁棒传输,随着丢包率的上升,DSFB-CS获得的重建性能缓慢下降,保证了面对不稳定信道时,解码端可以获得相对稳定的重构图像。
[Abstract]:Aim although the traditional joint source-channel coding scheme can achieve efficient compression performance, when the channel deterioration exceeds the error-correcting ability of channel coding, the performance of decoding reconstruction will decline sharply. In this paper, a robust SAR image coding transmission scheme is proposed by using the democracy of compression perception, and a series of methods are adopted to improve the rate-distortion performance of the scheme. Methods considering the rich edge information of SAR images, the direction lifting wavelet transform (DLWTT) with stronger directional representation ability is used to represent SAR images sparsely, and to eliminate the aliasing effect of recovering non-sparse signals in compression perception. The sparse filtering method is used to ensure the accurate recovery of large coefficients, and the efficient Bayesian reconstruction algorithm is used to obtain the high performance image reconstruction at the decoding end. Results compared with the traditional joint source-channel coding scheme (CCSDS-RS), the proposed scheme can achieve more robust coding transmission at the same bit rate, when the packet loss rate reaches 0. 05. The reconstruction performance of the proposed scheme DSFB-CS is obviously higher than that of CCSDS-RSs, and compared with the traditional scheme based on Bayesian reconstruction algorithm TSW-CS, the proposed scheme can improve the PSNR ratio to 3.9 dB. Conclusion the scheme DSFB-CS realizes the robust transmission of SAR images, and the reconstruction performance of DSFB-CS decreases slowly with the increase of packet loss rate, which ensures that the decoder can obtain relatively stable reconstructed images in the face of unstable channels.
【作者单位】：西安交通大学电子与信息工程学院;中国科学技术大学;
【基金】：国家自然科学基金项目(61373113) 中央高校基本科研业务费专项资金(xjj2012023)
【分类号】：TN957.52

【参考文献】