一种FPGA可实现的压缩传感采样方法  

FPGA Implementation Method of Compressed Sensing

作　　者：刘继忠[1] 金明亮[1] 马如远[2] 柴国钟[2] 王光辉[1] 

机构地区：[1]南昌大学机器人研究所,江西南昌330031 [2]浙江工业大学机械学院,浙江杭州310000

出　　处：《仪表技术与传感器》2013年第10期108-110,共3页Instrument Technique and Sensor

基　　金：国家自然科学基金(61273282);江西省教育厅自然科学基金(GJJ12005);横向合作基金(CX201206260185)

摘　　要：观测矩阵的构造一直是压缩采样硬件实现的难题,主要是因为符合压缩传感限制等容性原则的高斯矩阵和Noiselet矩阵目前在硬件上无法实现。文中基于FPGA小波稀疏,通过伪随机0、1序列完成观测矩阵构造和压缩传感采集,实现了一种适合硬件实现的伪随机序列采样机制的压缩传感采样方法。论文进行了FPGA伪随机序列矩阵构造实验验证,并通过matlab环境下的仿真和重构,表明了该种压缩传感采样方法的有效性。Because Gaussian matrix and Noiselet matrix which can satisfy the requirement of Restricted Isometry Property （RIP） are impracticable at present in hardware implementation for compressed sampling, the construction of observation matrix in hardware system has been a difficult problem in compressive sensing. Based on the wavelet sparsity, this paper realized a novel compressed sampling methodology by the help of pseudo-random sequence. The 0-1 pseudo-random sequence can be easily a- chieved on FPGA to construct observation matrix which are verified in the paper. Simulation experiments are carried out in Matlab, and the reconstruction result show that the pseudo-random sequence observation matrix has a good recovery ability and the method- ology is practical.

关 键 词：压缩传感 伪随机序列 FPGA 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]