一种基于编码孔径的压缩传感图像重构算法的改进  

Improved compressive sensing image reconstruction algorithm based on coded apertures

作　　者：张智诠[1] 丁晟[1,2] 

机构地区：[1]装甲兵工程学院控制工程系 [2]中国人民解放军73021部队

出　　处：《光学技术》2014年第2期133-139,共7页Optical Technique

摘　　要：针对压缩传感理论应用于实际系统成像时重构图像质量随图像采样率变化的问题,通过对正交匹配追踪算法进行改进,提出了一种利用空间光调制器实现编码孔径成像的压缩传感图像重构方法。该方法对传统的正交匹配追踪法迭代计算中已选入支撑集的列向量进行标记,并在下一次迭代计算中予以排除,从而减少了重构时间。在此基础上提出了将测量矩阵分别按行和按列排列进行重构后平均的图像增强算法。增强算法在达到同样重构质量时,减小了图像采样率,有利于图像数据的传输和存储。仿真实验验证了方法的有效性和稳定性,可为压缩传感技术的应用提供技术参考。Aiming at the problem of image reconstruction quality to change with sampling rate when compressive sensing applying in practical imaging system, with the improvement of orthogonal matching pursuit algorithm, a compressive sensing image reconstruction algorithm realized by spatial light modulator based on coded apertures is proposed. Compared to the traditional orthogonal matching pursuit algorithm, the column vectors which have been chosen in the support set are marked and are not engaged in the next iterative calculation process, thus the calculation time is reduced. The measurement matrix is arranged in row and column modes and used for image reconstruction respectively by the enhancement algorithm based on improved orthogonal matching pursuit algorithm for imaging. The final recovered image is the average image of two images recovered in row and column modes. The image sampling rate can be decreased when similar image qualities are obtained by the enhancement algorithm. The enhancement algorithm favours the transmission and storage of image data. The simulation experiment with several images shows effectiveness and stability of improved orthogonal matching pursuit algorithm for imaging and its enhancement algorithm, which provides a effective technical reference for CS application.

关 键 词：编码孔径 压缩传感 图像稀疏与重构 多元探测器 正交匹配追踪 

分 类 号：TN214[电子电信—物理电子学]