摘要
压缩感知理论中感知矩阵的约束等距性或累积互相关性对于信号重建有着重要意义。降低感知矩阵的累积互相关性或平均互相关性,能有效地提高信号重建表现。基于这个结论,文中提出了一种在给定稀疏表示矩阵的条件下,对投影矩阵进行迭代优化的方法。仿真结果表明,文中提出的优化方法能够很大程度上降低感知矩阵的累积互相关性和平均互相关性。同时实验进一步表明,文中提出的投影矩阵迭代优化方法明显提高了压缩感知正交匹配追踪算法的重建表现。
In compressive sensing theory,the restricted isometry property or cumulative mutual coherence plays an important role in signal reconstruction. Reducing the cumulative mutual coherence and the average mutual coherence of the sensing matrix can improve the reconstruction performance. Based on this conclusion,propose a novel iterative algorithm of optimization for projection matrix when the sparsity transformation matrix is deterministic. Simulation results show that the cumulative mutual coherence and the average mutual coherence of the obtained sensing matrix is reduced largely via the new algorithm. And the experiments also show that the new method proposed improves the reconstruction performance of OM P algorithm compared with optimizing before significantly.
出处
《计算机技术与发展》
2015年第3期95-98,共4页
Computer Technology and Development
基金
国家自然科学基金资助项目(60972041
60972045)
关键词
压缩感知
互相关性
投影优化
信号重建
compressive sensing
cumulative mutual coherence
projection optimization
signal reconstruction
