论文部分内容阅读
为实现具有任意稀疏结构的多量测向量(multiple measurement vectors,MMV)问题的快速重构.本文提出一种快速复数线性Bregman迭代算法(fast complex linearized Bregman iteration algorithm,FCLBI)并将其应用于逆合成孔径雷达(inverse synthetic aperture radar,ISAR)成像.首先,构建了任意稀疏结构的MMV信号模型并分析了其信号特征;其次,推导了复数条件下FCLBI算法的迭代公式用于重构MMV问题,将算法拓展到复数域使其更具普适性;然后,通过估计Bregman迭代过程中的停滞步长与感知矩阵优化相结合的方式减少迭代次数,从而可加快运算速度;最后,将算法应用于ISAR成像,进一步提高了稀疏恢复理论用于ISAR成像时的速度和抗噪性能.仿真和实测数据实验验证了算法的有效性.
To realize the fast reconstruction of multiple measurement vectors (MMV) with arbitrary sparse structure, a fast complex linearized Bregman iteration (FCLBI) algorithm is proposed and applied to inverse Firstly, the MMV signal model with arbitrary sparse structure is constructed and its signal characteristics are analyzed. Secondly, the iterative formula of FCLBI algorithm is derived for the reconstruction of MMV problem under complex conditions. The algorithm is extended to the complex number domain to make it more universal. Then, the number of iterations can be reduced by estimating the stagnation step length in the Bregman iterative process in combination with perceptual matrix optimization. Finally, the algorithm is applied to ISAR imaging further improves the speed and anti-noise performance of sparse recovery theory for ISAR imaging.Experimental results show that the proposed algorithm is effective.