基于多方向正交匹配追踪的压缩感知图像重构算法研究

针对压缩感知中图像信号的压缩比较大、重构效率较低的问题,提出一种基于正交匹配追踪的压缩感知图像重构算法。该算法在图像压缩阶段,利用测量矩阵对目标图像进行二次测量,进一步缩小压缩比。在图像重构阶段,将单次循环的正交匹配对象由一维向量拓展到二维矩阵,利用多方向正交投影对压缩数据进行图像重构。仿真实验结果表明,在缩小压缩比的情况下,可较准确地重构出原始图像;并且显著提高重构效率。

Compressed Sensing Reconstruction Algorithm for Image based on Multi-direction Orthogonal Matching Pursuit

The compressed sensing reconstruction algorithm for image based on multi-direction orthogonal matching pursuit was proposed to deal with the problem of the bigger compression ratio of image signal and lower efficiency of reconstruction algorihtm. In the stage of compression,the measurement matrix was used to measure the target image for two times.And the compression ratio was significantly reduced. In the stage of reconstruction, the orthogonal matching object was expanded from one dimensional vector to two-dimensional matrix in every circular, and the image was reconstructed by multi-directional orthogonal projection.The simulation results show that the image can be reconstructed accurately by the proposed algorithm when the compression ratio is reduced. And the efficiency of reconstruction is improved significantly.