基于压缩感知的MCSAR三维高分辨率快速成像研究

项目名称： 基于压缩感知的MCSAR三维高分辨率快速成像研究

项目编号： No.61501210

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 无线电电子学、电信技术

项目作者： 喻玲娟

作者单位： 江西理工大学

项目金额： 19万元

中文摘要： 多基线圆迹合成孔径雷达(MCSAR)模式下，雷达围绕被测目标进行多次不同高度的圆周飞行，不仅具有亚波长级的二维地距分辨率，而且具有很高的高度向分辨率。这在目标识别、高精度地形测绘、森林生物量估计、军事侦查等领域具有重大意义。然而，基于传统奈奎斯特采样定理的MCSAR三维成像，面临着数据采集难度大、数据量多，难以同时获得高的成像质量和成像效率问题。本项目拟将压缩感知理论应用到MCSAR三维高分辨率成像中，降低数据采集难度，以少量的数据实现三维目标的精确重建；将多核CPU-GPU异构平台应用于三维高分辨率成像处理的并行计算，在保证成像质量的同时，提高成像处理效率；开展可控实验，验证和完善基于压缩感知的MCSAR三维高分辨率快速成像的理论和处理方法。本项目的研究将为MCSAR三维成像系统的实用化奠定理论和技术基础。

中文关键词： 圆迹合成孔径雷达；高分辨率雷达成像；三维成像；压缩感知；并行计算

英文摘要： Multi-baseline Circular Synthetic Aperture Radar (MCSAR) is the imaging mode that radar observe objects over several circular trajectories with different height. It can obtain two-dimensional (2D) sub-wavelength resolution and high vertical resolution, which is very significant for target identification, high precision topographic mapping, estimation of forest biomass, reconnaissance, and so on. However, according to traditional Nyquist criteria, the data collection of MCSAR is highly complex, the imaging data is very big, and it is hard to obtain good image quality and high processing velocity at the same time. In this proposal, compressive sensing (CS) theory will be used in three-dimensional (3D) high-resolution MCSAR imaging, aiming to obtain low complexity of data measurement and reconstruct 3D objects with little data precisely. Multi-core CPU-GPU heterogeneous platforms is used in the parallel computing of 3D high-resolution imaging, aiming to obtain high processing velocity and keep good image quality at the same time. Controllable experiment will also be done to prove and improve the fast 3D high-resolution MCSAR imaging based on CS. This study will be to lay the theoretical and technical foundation for the utilization of the 3D MCSAR imaging system.

英文关键词： Circular Synthetic Aperture Radar;High-resolution radar imaging;Three-dimensional imaging;Compressive sensing;Parallel computing