近场稀疏MIMO面阵三维毫米波成像技术研究

Millimeter-wave (MMW) imaging technology has a widespread application prospect in the field of personnel surveillance. The current state-of-the-art in MMW imaging systems are constructed mainly by a one-dimensional (1-D) antenna array along with a mechanical scanning system to realize the three-dimensional (3-D) imaging by using a wideband signal. The main drawback of this type of systems is that the mechanical scanning slows down the imaging speed and improves the system unsteadiness. Against this problem, we will research the near-field 3-D MMW imaging technology with sparse MIMO area array via compressive sensing (CS). And the number of antenna elements and radio frequency channels can be further reduced compared with the traditional MIMO array. .Firstly, the optimization methods for designing the sparse MIMO area array is researched. The relationship between the random distribution of the antenna array and the RIP will be discussed. Secondly, we propose a multiple-frequency wavenumber domain 3-D MMW imaging algorithm for the cross-array MIMO structure. Due to no need of interpolation, the algorithm is faster than the traditional wavenumber domain algorithm and the back-projection algorithm, with much higher precision. Furthermore, this algorithm can be used in the subsequent CS image reconstruction. Finally, a large-scale CS image reconstruction algorithm is proposed based on using the measurement operator to solve the storage problem of the huge-scale measurement matrix existed in the 1-D optimization algorithms. .Through the aforementioned research, we hope to obtain the recovery guarantees of the CS-based sparse MIMO array imaging method. Also we will provide an optimization algorithm with high convergence rate based on the measurement operator. Finally, we hope this research can provide a theory basis for the application of the near-field 3-D MMW imaging.

毫米波成像技术在人体安检领域具有广阔应用前景，其主要基于一维天线阵列结合机械扫描及宽带信号实现三维成像检测。缺点是机械扫描增加了系统不稳定性，降低了成像速度。针对这一问题，拟研究基于压缩感知（CS）的稀疏MIMO面阵毫米波近场三维成像技术，与通常MIMO阵列相比，可明显降低阵元及射频通道数量。.首先，研究稀疏MIMO面阵优化设计方法，探讨天线阵列随机排布方式与约束等距性之间的规律。其次，针对十字阵提出一种多频波数域全息毫米波成像算法。该算法无需插值，与传统波数域算法及逆投影算法相比，其计算速度更快、精度更高，并且可用于后续CS图像重构中。最后，提出一种基于测量算子的大尺寸CS图像重构算法，以解决一维优化算法中测量矩阵尺寸过大的问题。.拟通过上述研究，获得稀疏MIMO阵列成像中的信号恢复条件，提出一种基于算子的快速优化算法，为基于CS的稀疏MIMO面阵系统在毫米波成像领域的应用提供理论基础。

毫米波成像技术在人体安检领域具有广阔应用前景，其主要基于一维天线阵列结合机械扫描及宽带信号实现三维成像检测。缺点是机械扫描增加了系统不稳定性，降低了成像速度。针对这一问题，本项目主要研究了基于压缩感知（CS）的稀疏MIMO面阵毫米波近场三维成像技术。.首先，研究了MIMO面阵优化设计方法，探讨了天线阵列随机排布方式与约束等距性之间的规律。其次，提出一种基于算子的大尺寸压缩感知成像方法。压缩感知的测量矩阵由多频波数域全息成像算法构建，解决了一维优化算法中测量矩阵尺寸过大的问题。第三，研究了多种阵面拓扑结构的MIMO成像技术，相比平面阵列，圆柱及折面形式的MIMO阵列与人体成像更匹配，可对目标区域提供更为均匀的天线波束覆盖，在更广范围内提高目标检测概率。第四，研究了近场单频点毫米波成像技术，提出一种基于多聚焦图像融合的三维成像方法。此外，提出一种两维相干因子，可以在两个维度有效降低成像旁瓣与鬼影目标。.通过上述研究，为基于MIMO面阵的毫米波成像技术在人体安检及无损检测领域的应用提供了理论基础与技术支撑。