智能系统学报Issue(1)：149-155,7.DOI:10.3969/j.issn.2013-0934.201309034
QR 分解与特征值优化观测矩阵的算法研究
An algorithm for measurement matrix based on QR decomposition and eigenvalue optimizatio
摘要
Abstract
Measurement matrix is a core part of compressed sensing .The column independence of measurement ma-trix and the incoherence between measurement matrix and sparse basis have a major impact on the quality of a re -constructed image .This paper proposes a new algorithm of measurement matrix based on QR decomposition and ei-genvalue.The column independence of the measurement matrix is increased by QR decomposition and at the same time the Gram matrix is optimized .Therefore , the eigenvalue of the normalized Gram matrix approximates to N/M so as to increases the incoherence between measurement matrix and sparse basis .The simulation results showed that the proposed method has excellent results on the aspects of increasing the quality of reconstructed image .In addi-tion, the stability of the reconstructed results had more apparent advantages than other algorithms in the case of less number of observed values.关键词
压缩感知/稀疏基/观测矩阵/重构算法/QR分解/特征值/列独立性/非相干性Key words
compressed sensing/sparse basis/measurement matrix/reconstruction algorithm/QR decomposition/eigenvalue/column independence/incoherenc分类
信息技术与安全科学引用本文复制引用
郑晓,薄华,孙强..QR 分解与特征值优化观测矩阵的算法研究[J].智能系统学报,2015,(1):149-155,7.基金项目
国家自然科学基金资助项目（61001140）． （）
