改进的压缩感知重构算法及其在图像融合中的应用  

Improved compressed sensing reconstruction algorithm and its application in image fusion

作　　者：李晖晖[1] 曾艳[1] 杨宁[1] 姚西文[1] 钱林弘 

机构地区：[1]西北工业大学自动化学院,陕西西安710129

出　　处：《南京理工大学学报》2014年第2期259-263,共5页Journal of Nanjing University of Science and Technology

基　　金：国家自然科学基金(60802084);西北工业大学基础研究基金(JC20110266)

摘　　要：为了提升压缩感知图像的重构精度,该文提出一种伪逆自适应压缩感知重构算法。该算法在总结已有贪婪算法的基础上,从最优原子的选择方式和支撑集更新过程两方面对已有算法进行改进。将算法应用于基于压缩感知原理的图像融合框架,实验结果表明,改进算法的重构图像质量优于基本贪婪类算法,应用于图像融合时可在较短的时间内得到更好的融合结果。In order to reconstruct images more accurate in compressed sensing,the pseudo-inverse adaptive matching pursuit（ PIAMP） reconstruction algorithm is put forward based on the existing greed algorithms. The algorithm modifies the existing algorithm from the selection method and support set updating process of the optimal atom. The algorithm is applied in the image fusion of the compressed sensing. The experimental results show that,the reconstruction effect resulting from PIAMP is better than that of other basic greed algorithms,and a good fusion result can be obtained in shorter time.

关 键 词：压缩感知 伪逆自适应算法 贪婪算法 图像融合 

分 类 号：TN911.73[电子电信—通信与信息系统]