分块压缩感知编码的重建图像改进算法

An improved reconstruction image algorithm based on sub-block compressed sensing coding

针对待编码图像分块实施压缩感知编码重建过程耗时较长问题,不是按照观测值与观测矩阵之间的关系来设计重建算法,而是在重建时先建立码书,然后直接从它中搜索出观测值意义下均方误差最小的最佳匹配块,作为重建图像子块.为了减少搜索范围,设置了剔除条件,设计出一个在码书中搜索最佳匹配块的限制搜索空间算法.四幅图像的仿真结果表明,重建图像质量对构成码书的原始图像不是特别敏感,具有一定的鲁棒性.它确实能够在重建图像质量有一定降低的情况下,其平均重建时间仅为正交匹配追踪算法的13.6%(16×16分块)与0.05%(32×32分块),为实时重建提供了一个较好的候选算法.

In order to solve the problem that the process of compressed sensing coding reconstruction takes a long time,the solution was not that the reconstruction algorithm was designed according to the relationship between the observed value and the observation matrix,but that the code book was first established during the reconstruction,and then the best matching block with the smallest mean square error under the meaning of the observed value was directly searched from it as the reconstructed image sub-block.To narrow down the search scope,exclusion criteria were introduced,and a limited search space algorithm was designed to search the best matching block in the code book.The simulation results of the four images showed that the quality of the reconstructed images was not particularly sensitive to the original images of the code book,and it had certain robustness.It was true that the average reconstruction time was only 13.6%(16×16 blocks) and 0.05%(32×32 blocks) of the orthogonal matching tracking algorithm when the quality of the reconstructed image was reduced to a certain extent,which provided a good candidate algorithm for real-time reconstruction.