采用正交多项匹配的块稀疏信号重构算法

发布时间：2018-04-10 23:33

本文选题：压缩感知 + 块稀疏信号　；参考：《信号处理》2014年06期

【摘要】：压缩感知,通过测量矩阵将原始信号从高维空间投影到低维空间,然后求解优化问题,从少量投影中重构出原始信号,是一种有效的信号采集技术。块稀疏信号是具有特殊结构的稀疏信号,其非零值是成块出现的。针对该信号的特点,提出一种采用正交多项匹配的块稀疏信号重构算法。该算法每次迭代选择多个最大相关子块,然后更新块索引集,以及迭代余量,最后求广义逆运算重构出原始信号。仿真结果表明,相比于大多数的现有算法,本文算法重构成功率较高,运行时间较短,复杂度较低。
[Abstract]:Compressed sensing, the original signal is projected from high-dimensional space to low-dimensional space by measuring matrix, then the optimization problem is solved, and the original signal is reconstructed from a small amount of projection. It is an effective signal acquisition technology.Block sparse signal is a sparse signal with special structure.According to the characteristics of the signal, a block sparse signal reconstruction algorithm using orthogonal multi-term matching is proposed.The algorithm selects multiple maximally correlated subblocks iteratively, then updates the block index set, and iterates the residue, and finally reconstructs the original signal by the generalized inverse operation.Simulation results show that compared with most existing algorithms, this algorithm has higher success rate, shorter running time and lower complexity.
【作者单位】：南京邮电大学通信与信息工程学院;
【基金】：江苏省自然科学基金(BK2011789) 东南大学毫米波国家重点实验室开放课题(K201318)资助课题
【分类号】：TN911.7

【相似文献】