一种合成孔径雷达参数化自聚焦的新方法

A Novel Parametric Autofocus Method for SAR Imaging

当合成孔径雷达(SAR)成像场景的低对比度下降时,现有SAR自聚焦方法的性能通常会随之下降甚至失效,并且由于缺乏“背景”的数学模型,自聚焦参数估计的极限性能通常是不能解析表达的。本文基于对成像场景中散射成分的分析,将SAR在短相干处理间隔中的慢变时间信号分解为目标、杂波和噪声三类独立的散射成分,得到了SAR慢变时间信号的一种参数化的统计模型。与现有的通过强散射成分的参数反演实现自聚焦的方法不同,本文通过估计模型杂波成分实现自聚焦,因此新方法尤其适用于低对比度场景。本文还推导了新方法的内在精度限制,以确定其在实际应用中的参数选择。最后,实际的SAR自聚焦结果证明了本文提出的新模型和新方法的有效性。

In synthetic aperture radar (SAR) imaging. low scene-contrast may deteriorate and even in-validate most of existing autofocus methods. And limit performance autofocus parameter estimation is alsodifficult to be verified analytically without appropriatc model for scene background.In this paper,by divid-ing the SAR slow-time signal into three isolated components(target· clutter and noise), a parametric statis-tical model is proposed in a coherent processing interval. Unlike the existing methods that realize autofocusbased on the prominent component’s parameters retrieval.a novel parametric autofocus method is proposedbased on the estimated parameters from the clutter.and it is especially applicable to low scene-contrastcase. Furthermore, the inherent accuracy limitation is obtained for the autofocus parameter estimation. Fi-nally, real measurement data is also exploited to investigate the effectiveness of the proposed model and thenovel method. The real autofocus results prove the model and the method to be effective.