matlab重建算法stomp,压缩感知图像重建算法的研究现状及其展望.pdf

压缩感知图像重建算法的研究现状及其展望.pdf

D陆33EE iE

γ，理想想

[本文献信息}李然，干宗良，崔子冠，等.压缩感知图像重建算法的研究现状及其展望[J]. 电视技术，20日，37 (19).

压缩感知图像重建算法的研究现状及其展塑

李然，干宗良，崔子冠，朱秀昌

(南京邮电大学江苏省图像处理与图像通信重点实验室，江苏南京 21C削3)

I 摘 要]压缩感知(Compressive Sensing ,CS) 理论是在已知信号具有稀疏性或可压缩性的条件下，对信号数据进行采集、编解码

的新理论。将压缩感知应用于图像压缩具有潜在的应用价值，压缩感知图像重建算法是该领域的热点问题。在对目前压缩感

知重建算法的文献进行分析和综合的基础上，首先阐述了压缩感知的基本原理及其各项关键技术，然后简要总结了当前流行的

压缩感知图像重建算法，给出了各种图像重建算法的仿真结果及分析，最后对影响压缩感知图像重建算法几个关键问题进行剖

析和展望。

[关键词]压缩感知;图像重建;稀疏;非相关测量

[中图分类号] TN91 1. 73 [文献标志码] A

Study Status and Pro叩ective of Compressive Sen血19 Image R配ons仕uction

U Ran, GAN Zongliang ,CUI Ziguan ,ZHU Xiuchang

(Jiangsu Pr，￥切M切'r咀即oviovi切归阳，;功t由阳阳FE四'e 旬句K tωa功bo佩'n lma咿茜伊'e Pr，斗o删吨&阳g伊'e 仙C01111阳阳ωω肌n厅， N叫i吨n唔占 UωM阳时r叫of Pos.ω臼 and η1'.'ele.阳t时ecoω01111押刷7

(A脑缸阳洒ct 1Compressive 缸Sensl圄囱s且m鸣I嘻g( CS)t由he回or哼y is a novel data collection and coding 由theor叮y under the condition that signal is sp缸'se or compressible. Image

compression base on CS h国 potential application value ,and the CS based image reconstruction algorithm is a hot topic. In the basis of many 1它lated litera-

tm酬，an overview on the state of the art of image reconstruction algorithms is given in this paper. Firsùy ,the CS foundations and its several key problems

are introduced. Then ,the current popular CS based image reconstruction algorithms are presented ,and 由e experiment results and analysis of th四e algo-

rithms are de配ribed. At the e时，the key problems in CS ba盟dima吕e reconstruction algorithm are summarized and prospected.

(Key words 1compressive sensi鸣;image reconstruction; spasity ;incoherent sampling

1 压缩感知简述 著系数编码，从而大大降低了编码复杂度。另外，压缩感知

的随机下采样操作甚至可在光域内直接进行(例如，单像素

近年来，压缩感知(Compressive Sensing , CS