基于压缩感知的图像鲁棒编码传输
2020-08-26王兰石熙张霖彭亚飞
王兰 石熙 张霖 彭亚飞


摘要:针对无线传感器网络这种资源受限的易丢包网络中图像鲁棒传输问题,本文提出了基于压缩感知的图像鲁棒编码传输算法,实现单个操作中同时采样压缩并且能够进行鲁棒传输。采用小波变换(DWT)对图像进行稀疏表示,再对稀疏系数进行块Hadamard压缩感知测量,为获得图像的高性能重建,在解码端采用高效的CRSR重建算法。通过实验仿真表明,密文在传输过程中能够抵抗高丢包率的攻击,随着丢包率的上升,本算法获得的重建性能缓慢下降,在丢包率为0.3时,解码端仍可以获得相对稳定的重构图像,说明该算法具有抵抗丢包攻击的鲁棒性,能够满足图像鲁棒传输。
关键词:压缩感知;鲁棒编码;丢包攻击
中图分类号:TP309 文献标识码:A
文章编号:1009-3044(2020)20-0017-03
Robust Coding Transmission for Image Based on Compressive Sensing
WANG Lan. SHI Xi. ZHANG Lin. PEING Ya-fei
(School of Mathematics and Information Engineering, Chongqing University of Education, Chongqing 400147, China)
Abstract: Image robust transmission in a wireless visual sensor network (WVSN) with limited resources over an unreliable and easyto packet loss wireless channel is challenging. A robust image coding algorithm based on compressed sensing is designed in this pa-per, which achieves simultaneous sampling and compression in a single operation and can perform robust transmission. The imagesare sparsely represented with wavelet transform (DWT) and sampled with block Hadamard pseudo-random measurement matrix.Then the measurements are quantized and packed for transmission to the decoding end. We adopt an efficient GRSR reconstructionalgorithm to attain high-performance image reconstruction at the decoding end. The experimental result shows that ciphertext canresist high packet loss attacks during transmission. The reconstruction performance obtained by this algorithm slowly decreaseswiththe packet loss rate increases. The decoder can still obtain the relatively stable reconstructed image when the packet loss rate is0.3. The result shows that the algorithm is robust against packet loss attacks and can meet the robust transmission of images.
Key words: compressed sensing; robust coding; packet loss attacks
無线传感器网络包含低成本和低功耗的传感器节点,由于信道能量和带宽资源有限,传输的图像数据需要被压缩。由于传感器节点具有低功耗特点,无法持续在图像压缩和图像传输中进行复杂计算,所以需要在图像编码和传输算法中找到传感器节点的计算复杂度和压缩性能之间的平衡点。传统的HEVC静态图像压缩和JPEG2000压缩方法的良好压缩性是以高复杂计算为代价,所以并不适合无线传感器网络。同时,图像编码和传输方案应该考虑无线传感器网络中的分组丢失问题,若重要的分组丢失,那么解码器就不能正确地解码出原图像。
压缩感知理论( Compressed Sensing,cs)表明可以从欠采样测量重建可压缩信号,所以CS可应用于无线传感器网络中的友好的图像压缩。……
