多测量向量块稀疏信号重构ISAR成像算法

为实现有限脉冲快速逆合成孔径雷达(inverse synthetic aperture radar,ISAR)稀疏成像,利用ISAR目标块状结构特征,提出一种基于多量测向量(multiple measurement vectors,MMV)模型的块稀疏信号重构ISAR成像算法。首先,构建MMV稀疏成像模型,将ISAR成像转化为MMV块L₀范数的稀疏重构问题。其次,选用负指数函数序列作为平滑函数去近似块L₀范数,通过构建一个递减的参数序列,对平滑函数优化求解,采用梯度投影方法将所求解投影到可行解空间。最后,增加修正步骤,确保沿着最速下降方向对块稀疏信号优化求解。仿真结果验证了本文算法在成像时间和成像质量方面具有优势。

Multiple measurement vectors block sparse signal recovery ISAR imaging algorithm

In order to obtain fast inverse synthetic aperture radar (ISAR) sparse images with finite pulse, a multiple measurement vectors (MMV) model block sparse signal recovery ISAR imaging algorithm is proposed by utilizing the block structure of targets. Firstly, the MMV sparse imaging model is established, ISAR imaging is converted into the MMV block sparse signal recovery problem. Then, one negative exponential function sequence is used as the smoothed function to approach the block L₀ norm, the optimization solution of the smoothed function is obtained by constructing a decreasing sequence, the solution is projected into the feasible set by the gradient projection algorithm. Finally, the revised step is added to ensure the searching direction of the optimization value of the block sparse signal is the steepest descent gradient direction. Simulation results verify the proposed algorithm has advantages in imaging time and imaging quality.