基于压缩感知理论的PIE显微成像研究  

Microscopic PIE Imaging With Theory of Compressive Sensing

作　　者：何靖[1] 刘诚[1] 高淑梅[1] 王继成[1] 王跃科[1] 朱健强 

机构地区：[1]江南大学理学院光信息科学与技术系,江苏无锡214122 [2]中国科院上海光学精密机械研究所,上海201800

出　　处：《光学学报》2014年第5期91-96,共6页Acta Optica Sinica

基　　金：江苏省自然科学基金(BK2012548);江南大学自主科研基金(JUSRP211A20)

摘　　要：为了克服PIE成像中所面临的数据量过大的问题,将压缩感知理论用于PIE成像。将采样到的衍射斑稀疏变换并压缩后,可以显著减少需要存贮的数据量。再现过程中选用子空间匹配追踪算法(SP)或者正交匹配追踪算法(OMP)重构出散射斑的原始分布,用常规的PIE算法进行图像重建。模拟和实验结果均表明,当压缩采样率在30%的时候就能重构出很好的图像。和OMP重构算法相比,SP算法更适合在PIE成像中应用。In order to overcome the problem of large amounts of data in PIE imaging, the theory of compressive sensing is adopted for PIE imaging. Diffraction patterns are sparsified and compressed, which can reduce the amount of data requiring save. The original distribution of scattering spot can be reconstructed with subspace persuit （SP） or orthogonal matching pursuit （OMP） tuning algorithm in the retrieval process, then image reconstruction is realized using regular PIE algorithm. Simulation and experimental results show that the images can be reconstructed perfectly when the compression ratio is 30 %. SP algorithm is more appropriate than OMP algorithm for PIE imaging.

关 键 词：图像处理 PIE成像 稀疏 压缩感知 重构算法 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]