基于融合压缩感知的Landweber算法研究

Compressed Sensing-Fused Landweber Algorithm

为提高Landweber算法的收敛速度,将压缩感知理论引入Landweber算法中,理论上证明了该算法的可行性并用改进的距离驱动进行系统矩阵的建模。实验结果证明:融合算法进行图像重建时,算法的收敛速度及采样时间均优于Landweber算法。并用GPU对重建算法进行加速,大幅度提高了算法的执行速度。

In order to increase the convergence speed of Landweber algorithm, compressed seasoning is combined in Landweber algorithm. Its feasibility is discussed theoretically and improved distance driven method is used to generate system matrix. The simulation experiment proves that fusion algorithm is better than traditional Landweber algorithm in convergence speed and time of getting projection data. What＇s more, fusion algorithm based on GPU can accelerate the execute time of the algorithm.