基于图拉普拉斯嵌入的合成孔径雷达时变窄带干扰抑制算法

合成孔径雷达(synthetic aperture radar, SAR)作为一种宽带系统, 常与同频段内其他有源电磁系统发生冲突, 这些信号相较于宽带SAR系统而言, 大部分是窄带干扰(narrow-band interference, NBI), 会对高分辨SAR成像系统产生严重干扰。早期关于NBI抑制问题的研究中, 很少有人注意到NBIs在不同脉冲之间可能具有局部时变特性, 这削弱了某些经典方法在干扰抑制上的性能。因此, 本文提出了一种图拉普拉斯嵌入(graph Laplacian embedding, GLE)算法, 通过在不同脉冲信号之间构建拉普拉斯嵌入关系来抑制NBIs。这使得局部时变的干扰能够被嵌入到非线性低维流形中, 并被有效去除。对实测受NBI干扰的SAR数据进行处理, 处理后的结果证明了所提方法的有效性。

Time-varying narrow-band interference suppression algorithm for SAR based on graph Laplacian embedding

As a wideband radar system, the synthetic aperture radar (SAR) usually conflicts with other electromagnetic systems working in the same frequency bands. These signals, most of which are narrow-band interference (NBI) for SAR systems, severely interfere SAR systems to generate a high-resolution image. Many previous research results focused on the NBI suppression problem, however, few of them noticed that the NBIs may be time-varying in a local region among different pulses, which restricts the suppression performance of some classic methods. Hence, in this paper, a graph Laplacian embedding (GLE) algorithm is proposed to suppress NBIs by constructing the Laplacian embedding connections between different pulses of signals. As a result, locally time-varying interferences are embedded in a non-linear low-dimensional manifold and can be effectively mitigated. The real SAR data measured with NBI are processed, and the results show the efficiency of the proposed algorithm.