一种确定性测量矩阵与快速恢复算法  

Deterministic Measurement Matrix and Fast Recovery Algorithm

作　　者：唐川雁 朱皓 TANG Chuan-yan;ZHU Hao

机构地区：[1]杭州电子科技大学通信工程学院,浙江杭州310018

出　　处：《通信技术》2018年第3期550-555,共6页Communications Technology

摘　　要：为了解决压缩感知中传统的测量矩阵不确定性大、硬件实现复杂等问题,提出了一种简单的二元块对角确定性测量矩阵,同时在离散余弦变换域中使用阈值法来控制信号的稀疏水平,并在此基础上提出了一种快速恢复算法。实验表明,与传统的测量矩阵相比,所提测量矩阵可以有效降低硬件上的实现难度,并得到更好的重构质量,同时所提快速恢复算法与OMP、St OMP算法相比,在信噪比和恢复时间上均具有更好的性能。In order to solve the problem of large uncertainty and complex hardware implementation in traditional measurement matrix for compressed sensing,a simple binary block diagonal deterministic measurement matrix is proposed.Meanwhile,the threshold method is employed to control the sparse level of the signal in the discrete cosine transform domain,and on this basis,a fast recovery algorithm is proposed.The experiment shows that compared with the traditional measurement matrix,the proposed measurement matrix can effectively reduce the difficulty of hardware implementation and acquire better reconstruction quality.In addition,the proposed fast recovery algorithm has better performance than the OMP and StOMP algorithms in SNR and recovery time.

关 键 词：压缩感知 确定性测量矩阵 离散余弦变换域 恢复算法 

分 类 号：TN911.7[电子电信—通信与信息系统]