合成孔径雷达自回归线性预测带宽外推超分辨率成像算法

SAR superresolution imaging method based on 2D linear autoregressive prediction bandwidth extrapolation

提出一种合成孔径雷达(SAR)二维自回归线性预测带宽外推超分辨率成像算法。根据SAR成像机理推导SAR线性预测带宽外推的信号模型;利用观测数据和SAR信号模型,估计表征信号带宽内和带宽外数据性质的统计特性,计算二维自回归(AR)模型参数;采用层迭方式外推观测数据,生成高分辨率图像。分别利用距离向数据、点目标仿真数据和实测数据对本文方法进行有效性验证。研究结果表明:本文方法可以在不增加硬件系统负担的情况下,提高影像分辨率。

A synthetic aperture radar （SAR） bandwidth extrapolation method was presented to improve SAR image resolution using two-dimensional auto regressive （AR） linear prediction extrapolation. The signal model was derived for linear prediction bandwidth extrapolation. The method used the signal model and the obtained data as the prior information to estimate the 2D AR parameters. A lap-to-lap approach was developed for the efficient bandwidth extrapolation to obtain a large dimension spectrum in frequency domain. The proposed method can be proved effective using experiments of range data, dot simulation data and real SAR data. The results show that paper＇s method can improve the image resolution without increasing the SAR hardware.