采用压缩感知理论的穿墙稀疏成像恢复算法需要感知矩阵满足有限等距性质(RIP),最直接的验证方法是判定感知矩阵的相干系数是否较小。针对现有的穿墙多输入多输出(MIMO)阵列稀疏成像方法没有验证感知矩阵是否满足RIP 性质而容易出现重构失败并导致成像模糊的问题,提出一种联合感知矩阵优化的穿墙MIMO阵列稀疏成像方法。该方法首先依据配置指标将阵元配置为两端发中间收和分时复用的模式,既能使感知矩阵的相干系数最小,又能获得均匀而不冗余的等效虚拟阵元;然后,从中选取部分能够满足感知矩阵相干系数最小的虚拟阵元组合,使用可分离逼近结构稀疏恢复算法充分考虑扩展目标信号的结构稀疏先验信息对其进行稀疏成像重构;最后,选取成像性能指标较好的一组作为成像结果。仿真和实验结果表明,该方法降低了运算量和虚拟阵元间的干扰,节约了硬件成本,提高了算法的稀疏重构性能,获得了高分辨的穿墙扩展目标成像。

The compressed sensing sparse imaging restoration algorithm about through-wall radar requires that the sensing matrix satisfies the RIP property, and the most direct verification method is to determine whether the coherence coefficient of the sensing matrix is small. The existing sparse imaging methods of MIMO through-wall array do not verify whether the sensing matrix satisfies the RIP properties and prone to cause reconstruction failure and blurred images. In this paper, a joint sensing matrix optimization sparse imaging method for through-wall MIMO array is proposed. Firstly, the array element is configured as the time division multi-plexing mode that can be transmitted at both ends and received in the middle according to the configuration index requirements, which can minimize the coherence coefficient of the sensing matrix and obtain uniform and irredundant equivalent virtual array. And then a combination that have some virtual array elements which can satisfy the minimum coherence coefficient of the sensing matrix is selected. The structure sparse compressed sensing recovery algorithm based on group separable approximation is used to reconstruct sparse image, which takes full account of the sparse prior information of extended target signal. Finally, a group of better imaging performance indexes is selected as the imaging result. The simulation and experimental results show that this method reduces the computational complexity and the interference between the virtual array elements, saves the hardware cost, improves the sparse reconstruction performance of the algorithm, and achieves high resolution through-wall extended target imaging.

TN957.51

国家自然科学基金资助项目(61461012);广西无线宽带通信与信号处理重点实验室项目(GXKL06160106); 广西自然科学基金资助项目(2017GXNSFAA198050)