摘要
为提升压缩感知信号重构效果,提出一种改进的小波变换子空间匹配追踪图像重构。根据图像信号小波变换后各子带系数的不同特性,保留小波变换低频子带系数,滤除、筛选高频子带系数,从而更有效地提取小波变换后图像的显著信息。然后基于压缩感知理论对处理后的高频子带系数进行随机观测,并结合子空间匹配追踪算法完成恢复计算,最后与低频子带系数结合实现图像的重构。仿真实验验证了该算法可以有效地提高图像信号的重构效果。
In order to improve the quality of signal reconstruction based on Compressed Sensing （CS） theory,this paper proposes an image reconstruction approach using Match Pursuit method on wavelet sub-space. According to the different characteristics of four wavelet sub-bands, the method proposed preserves the low-pass sub-band coefficients and filters high-pass coefficients to extract salient information more efficiently. Based on the CS theory,the filtered high-pass wavelet coefficients are measured with a Gaussian distributed matrix and then reconstructed by the Subspace Match Pursuit method. The reconstructed high-pass coefficients are then combined with the preserved low-pass coefficients to obtain the reconstructed image. The simulation results show that the quality of reconstructed image is significantly improved by this method.
出处
《火力与指挥控制》
CSCD
北大核心
2013年第10期34-37,共4页
Fire Control & Command Control
基金
国家自然科学基金(61071170)
教育部新世纪优秀人才支持计划资助项目
关键词
压缩感知
高频子带
匹配追踪
信号重构
compressed sensing, high-pass sub-band, match pursuit, signal reconstruction
