摘要
针对压缩感知理论仿真实验在处理较大图像数据时,传统的一维压缩感知算法不实用、无法处理数组爆炸从而导致数据无法处理的问题,提出一种改进的分块存储按列恢复的分块压缩感知方法,并根据等效RIP性质加入传感矩阵的不相干性判断。实验结果表明,该算法能很好地解决传统一维信号的压缩感知处理方法无法处理二维信号的问题,相比于简单的分块压缩感知算法更为有效、算法运行时间更低、内存消耗更小、图像恢复精度更好。
In the case of simulation experiments for compression sensing theory when dealing with a larger image data,the traditional one-dimensional compression sensing algorithm is not practical,unable to cope with an array which can lead to the problem of failing to process data.This paper puts forward an improvedblock compression sensing method based on block storagerecovery by column.And according the nature of the equivalent RIP,this method joins incoherence of sensing matrix judgment.The experimental results show that the proposed algorithm can solve the problem that the traditional one-dimensional compressed sensing algorithm cannot handle two-dimensional signals,and is more effective than the simple block-compressed sensing algorithm.The algorithm takes less time,consumes less memory,and improves image recovery accuracy.
作者
黄超
吴亚娟
Huang Chao;Wu Yajuan(China West Normal University,Nanchong 637009,China)
出处
《廊坊师范学院学报（自然科学版）》
2020年第4期21-25,共5页
Journal of Langfang Normal University(Natural Science Edition)
基金
西华师范大学英才科研基金项目(17YC163)。
关键词
压缩感知
分块处理
图像恢复
compressed sensing
block processing
image restoration
