CAJ | 학술논문

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。
위료제승압축감지도상적중구정도，해문제출일충위역자괄응압축감지중구산법。해산법재총결이유탐람산법적기출상，종최우원자적선택방식화지탱집경신과정량방면대이유산법진행개진。장산법응용우기우압축감지원리적도상융합광가，실험결과표명，개진산법적중구도상질량우우기본탐람류산법，응용우도상융합시가재교단적시간내득도경호적융합결과。
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.