基于准循环低密度奇偶校验码的压缩感知测量矩阵

发布时间：2018-07-06 15:42

本文选题：压缩感知 + 随机测量矩阵　；参考：《计算机应用》2014年11期

【摘要】：针对随机测量矩阵元素随机产生、不易于硬件实现的缺点,利用有限域上准循环低密度奇偶校验(QCLDPC)码奇偶校验矩阵的构造方法,设计了一种确定性的结构化稀疏测量矩阵。由于QC-LDPC码的信道编解码性能较好,故以此为基础构造压缩感知(CS)测量矩阵预计有较好的性能。分别用一维和二维信号的CS重建实验验证新矩阵的性能,结果表明,与常用的测量矩阵相比,在相同的重建算法和压缩比条件下,新矩阵对应的重建误差较低,在峰值信噪比(PSNR)的评价指标上有所提高(0.5~1 dB)。特别地,所提的确定性测量矩阵在结构上具有对称特性和准循环特性,如将其应用于硬件实现,可降低物理内存的需求量与硬件实现的复杂度。
[Abstract]:Aiming at the disadvantage that random measurement matrix elements are generated randomly and are difficult to be implemented in hardware, a deterministic structured sparse measurement matrix is designed by using the construction method of quasi cyclic low density parity check (QCLDPC) code parity check matrix in finite field. Because QC-LDPC codes have good performance in channel coding and decoding, the performance of CS measurement matrix based on QC-LDPC codes is expected to be better. The performance of the new matrix is verified by the CS reconstruction experiments of one and two dimensional signals respectively. The results show that the reconstruction error of the new matrix is lower than that of the common measurement matrix under the same reconstruction algorithm and compression ratio. The peak signal-to-noise ratio (PSNR) was improved (0.5dB). In particular, the proposed deterministic measurement matrix has symmetrical and quasi-cyclic characteristics in structure. If it is applied to hardware implementation, the requirement of physical memory and the complexity of hardware implementation can be reduced.
【作者单位】：南通大学电子信息学院;
【基金】：国家自然科学基金资助项目(61171077) 交通部应用基础研究项目(2011-319-813-510) 南通大学研究生科技创新计划项目(YKC13003) 南通大学创新人才基金资助项目(2009)
【分类号】：TN911.7

【参考文献】