基于混合采样的压缩感知重构算法

为了提高图像的重构质量和缩短重构时间,同时保持较高的压缩比,提出了一种基于混合采样的压缩感知重构算法。将图像划分为感兴趣区域和非感兴趣区域,对感兴趣区域采用恢复质量较好的正交匹配追踪算法,对非感兴趣区域采用恢复时间较短的分段正交匹配追踪算法。感兴趣区域图像中除感兴趣区域外,其他部分灰度置零以增加采样率和图像稀疏度。实验表明,该方法可以较好恢复图像感兴趣的区域,并保持较高压缩比。

Compressed Sensing Reconstruction Algorithm Based on Hybrid Sampling

In order to improve the image reconstruction quality and the reconstruction time with high compression ratio, a compression perception reconstruction algorithm based on hybrid sampling is proposed. The image is divided into the interested and non-interested area. It used the orthogonal matching pursuit (OMP) algorithm with better quality of recovery for the interested area, and the Stagewise Orthogonal Matching Pursuit algorithm with the recovery time shorter for the non-interested area. In the interested image, the gray level of the parts except the interested region is set to zero, so as to increase the rate of sampling and image sparse degree. Experiments show that this method could restore the interested area of an imagine better, and maintain a high compression ratio.