基于选择性测量的压缩感知去噪重构算法

发布时间：2018-01-07 17:25

本文关键词：基于选择性测量的压缩感知去噪重构算法　出处：《通信学报》2017年02期 　论文类型：期刊论文

【摘要】：针对压缩感知中噪声折叠现象严重影响稀疏信号重构性能的问题,提出一种基于选择性测量的压缩感知去噪重构算法。首先从理论上解释了压缩感知中噪声折叠现象;然后提出一种基于测量数据的特征统计量,推导分析其概率密度函数,并基于此构造一种噪声滤波矩阵,用于优化测量矩阵,实现智能地选择信号分量、过滤噪声分量,提高测量数据信噪比;最后,通过增加测量数据获取次数可进一步提升算法重构性能。仿真实验表明,基于选择性测量的压缩感知去噪重构算法明显改善了低信噪比条件下信号的重构性能。
[Abstract]:For the compressed sensing noise in the folding phenomenon seriously affects the sparse signal reconstruction performance problems, put forward a kind of selective measurement of compressed sensing reconstruction algorithm based on denoising. First theoretically explains the compressed sensing noise folding phenomenon; and then a statistic analysis based on measured data, the probability density function, and based on this to construct a noise filter matrix is used to optimize the measurement matrix, to realize the intelligent selection of signal components, filtering noise components, improve the measurement data signal-to-noise ratio; finally, by increasing the number of measurement data acquisition will further enhance the performance of reconstruction algorithm. Simulation results show that the selective measurement of compressed sensing reconstruction denoising algorithm can obviously improve the low SNR than the reconstruction performance conditions based on the signal.

【作者单位】：解放军信息工程大学;
【基金】：国家自然科学基金资助项目(No.61401511)~~
【分类号】：TN911.7
【正文快照】： 1引言随着信息技术的飞速发展,人们对信息量的需求越来越大,携带信息的信号带宽也越来越宽,以传统奈奎斯特采样定理为基础的信号处理框架对采样速率的要求也越来越高,有时甚至高达数吉赫兹,而现有硬件设备尚无法满足其对模数转换能力的要求[1]。压缩感知[2~4](CS,compressiv

【参考文献】