基于OMP算法的图像重构研究与FPGA实现

Study of Image Reconstruction Based on OMP Algorithm and FPGA Implementation

摘要:

针对高分辨率的图像在采集过程中存在数据量较大的问题,提出了一种基于正交匹配追踪(OMP)算法的图像重构方法,设计了OMP算法的硬件结构,并在FPGA平台上进行了仿真验证；首先,研究了压缩感知算法的基本原理；然后,分别基于匹配追踪算法(MP)和正交匹配追踪算法实现了图像的重构；最后,通过仿真对比分析了这两种方法的图像重构结果,OMP算法误差在10-15量级,明显优于MP算法的10-3误差量级,并且OMP算法的迭代收敛性也优于MP算法。

Abstract:

For the high-resolution image need a large quantity of data in an acquisition process, a method based on orthogonal matching pursuit (OMP) algorithm for image reconstruction has been proposed and verified through the FPGA platform. Firstly, the basic principle of compressed sensing algorithm has been studied. Secondly, the image reconstruction can be realized based on the matching pursuit algorithm (MP) and orthogonal matching pursuit algorithm respectively. Lastly, compared with the simulation results of image reconstruction based on the above two methods, show that the OMP algorithm is superior to MP algorithm not only at the convergence but also the reconstruction effect.