利用压缩感知方法的高分辨率三维层析SAR研究

High Resolution 3D Tomographic SAR with Compressive Sensing

利用压缩感知方法解决SAR三维层析成像中的频谱估计问题,弥补了传统谱估计方法的一些缺陷,准确重建了来自不同散射目标的回波信号。实验结果表明,该方法不仅能够有效地区分影像中同一像元中的多个散射体,而且可以较为准确地获取目标散射体相对强度分布及高程信息,为目标的分类和识别提供更多依据,显现出对数据量要求较少及抗噪声性能强等优势。

In 2D SAR imaging,the ambiguity problem is hindering the SAR image interpretation.A new method of compressive sensing（CS） is used to solve the spectral analysis problem in SAR tomography for the first time.The signal reconstruction is demonstrated using simulations and TerraSAR-X spotlight data of Berlin.TomoSAR with CS has huge potential,because not only multiple scatterers inside one resolution cell can be separated,the elevation positions and height of scatters can be accurately measured in the reflectivity slice,but also CS is more robust to noise,has lower computational effort and use less data.