Abstract:

Focusing on the problems, including complexity of computation and the sensitivity to noise, of the traditional Compressive Sensing based TWRI（Through-the-Wall Radar Imaging）algorithms, an Inexact Augmented Lagrange Multiplier（IALM） based on the compressive sensing framework is proposed for TWRI. Considering the low rank of wall echo signal and the sparsity of target signal, this paper regards TWRI as the optimization problem of the regularized least square optimization problem, and is transformed into a compound optimization problem with kernel-norm and [l1]-norm. The Inexact Augmented Lagrange Multiplieralgorithm（IALM） is used to update the wall clutter matrix and the target matrix alternately, so as to complete the target reconstruction. The simulation results show that the target to clutter ratio provided by the proposed algorithm is better and significantly improves the target imaging accuracy and the processing speed.

Key words: compressive sensing, Through-the-Wall Radar Imaging（TWRI）, Inexact Augmented Lagrange Multiplier（IALM）

摘要：

针对传统压缩感知穿墙雷达成像算法存在的计算过程复杂、对噪声较敏感等问题，提出一种压缩感知框架下的基于非精确增广拉格朗日乘子（Inexact Augmented Lagrange Multiplier，IALM）的穿墙成像算法。考虑到墙体回波信号的低秩性和目标信号的稀疏性，将穿墙成像视为正则化最小二乘优化问题，并转化为一个含核范数和[l1]范数的复合优化问题。通过非精确增广拉格朗日乘子法交替更新迭代求解墙杂波矩阵和目标矩阵，从而完成目标重构。仿真结果表明所提算法的信杂比较好，并在显著提高目标成像精度的同时提高了处理速度。

关键词: 压缩感知, 穿墙雷达成像, 非精确增广拉格朗日乘子