基于 Kronecker 压缩感知的宽带 MIMO 雷达高分辨三维成像

摘要

In the Three Dimension (3D) imaging using a wideband Multiple-Input Multiple-Output (MIMO) radar, the resolution in the two cross-range dimensions is usually not satisfactory in practice, limited by the length of the MIMO radar array. In the paper, the Compressive Sensing (CS) theory is applied to realize the super resolution in the two cross-range dimensions. Firstly, a joint two dimensions super resolution method via Kronecker CS (KCS) is proposed, to avoid losing the coupling information among different dimensions, which will happen when the super resolution is just considered in each dimension separately. Then, in order to solve the problem of huge storing and computing burden in KCS, a dimension reduction method is proposed further by utilizing the prior information of the low resolution 3D image. Finally, the validity of the method is verified with simulated data and real measured data experiments.%在宽带多输入多输出(MIMO)雷达3维成像中，MIMO 雷达收发阵元数量和空间分布的限制会导致图像的2维横向分辨率难以满足实际需求。该文利用压缩感知(CS)理论来实现图像在2维横向上的超分辨。考虑到对信号的每一维分别进行超分辨会损失各维间的耦合信息，提出一种基于 Kronecker CS(KCS)的2维联合超分辨方法；为解决 KCS 在多维高分辨应用中存储量大、计算效率低的问题，进一步提出了一种基于低分辨3维图像先验信息的降维 KCS 方法。仿真和实测数据实验验证了方法的有效性。