基于压缩感知理论的近场三维毫米波全息成像方法研究

项目名称： 基于压缩感知理论的近场三维毫米波全息成像方法研究

项目编号： No.61301191

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 无线电电子学、电信技术

项目作者： 李世勇

作者单位： 北京理工大学

项目金额： 25万元

中文摘要： 三维毫米波全息成像技术在人体隐匿目标检测、医学成像等领域具有广阔应用前景。传统方式是采用一维天线阵列结合机械扫描，及宽带信号实现三维全息成像。其缺点是机械扫描会降低成像速度，若采用两维天线阵列，则可以实现快速甚至实时成像，但两维阵列的天线单元数目巨大，会增加系统复杂度和研制成本。 因此，拟研究基于压缩感知（CS）理论的近场三维毫米波全息成像方法。毫米波图像在一定变换基下具有稀疏性，由CS理论可知，利用远少于Nyquist采样情况下的测量数据即可实现图像的精确重建。首先，研究毫米波图像稀疏表示方法，及两维随机稀疏天线阵列设计方法。其次，研究图像的多维优化重建算法。由于测量矩阵相对图像的三维方向是不可分的，所以拟研究基于测量算子的图像重建算法，以避免测量矩阵过大的问题。另外，拟采用非均匀快速傅里叶变换（NUFFT）替代FFT与插值，以进一步提高计算精度和计算速度。

中文关键词： 毫米波成像；压缩感知；全息成像；稀疏阵列；优化算法

英文摘要： Three-dimensional (3-D) millimeter-wave (MMW) holographic imaging technology has wide application prospects in concealed object detection for humman body security inspection, medical imaging, and etc. The traditional approach for 3-D MMW imaging was based on utilizing 1-D antenna array scanned along its normal direction and wide bandwidth signal. However, the mechanical scanning mode slows down the imaging speed. If 2-D antenna array is used, rapid even real time imaging can be achieved. But the number of antennas will be huge, then increasing the system complexity and the development cost. Therefore, we will research the near-field 3-D MMW holographic imaging method via compressive sensing (CS) theory. The MMW images have sparse representations in some proper basis, thus, based on CS, the images can be reconstructed exactly from far fewer measured data compared with the traditional methods according to the Nyquist theorem. Firstly, we research the sparse representations of MMW images and the design methods for 2-D randomly distributed sparse antenna array. Then, the multi-dimensional optimization methods for reconstruction of MMW images are researched. Due to the fact that the three dimensions of measurement matrix are inseparable for the three directions of MMW images, we will research the image reconstruction

英文关键词： Millimeter-wave imaging；Compressive sensing；Holographic imaging；Sparse array；Optimization algorithm