基于自适应子空间追踪的层析SAR成像算法
2014-12-18任健刘梅
任健+刘梅
摘 要: 在对分布式SAR进行数据降采样下会信号的三维处理增加不少难题。其中在解决频域距离弯曲校正时,由于方位向的降采样使数据不再满足奈奎斯特定理,导致在多普勒域计算距离偏移量时会出现数据的混叠。针对该问题,提出了基于LMS估计的距离弯曲校正算法,该方法根据最小均方估计思想估计权值系数完成方位向的插值,有效解决了该条件下的距离弯曲问题。针对高层成像中稀疏阵列导致基线数量有限且不均导致成像分辨率差的问题,提出了基于压缩感知的自适应子空间追踪方法来提高高度维成像的分辨性能,相比于正交匹配追踪算法,它能实现对迭代得到候选解的同步检验,避免了错误结果积累的问题,有效提高了成像的质量。
关键词: 稀疏数据; 距离弯曲校正; 层析; 压缩感知
中图分类号: TN957.52?34 文献标识码: A 文章编号: 1004?373X(2014)24?0072?04
SAR tomography algorithm based on adaptive subspace tracking algorithm
REN Jian, LIU Mei
(Research Institute of Electronic Engineering Technology, Harbin Institute of Technology, Harbin 150001, China)
Abstract: The undersampling for distributed SAR system may cause many problems in 3D processing of the signal. When dealing with the range cell curve correction, due to the down?sampling in azimuth direction, the azimuth data are no longer satisfied the Nyquist theorem, which may cause the data aliasing when the range offset is calculated in the Doppler domain. To solve the problem, a new method named range cell curve correction base on LMS estimation is presented. Benefited from the idea of LMS estimation, this method completes azimuth interpolation process, and the range cell curve problem can be solved under this condition. Moreover, aiming at the low resolution caused by the sparse array for SAR tomography, a new method named adaptive subspace tracking algorithm with compressive sensing is presented to improve the imaging resolution. Compared with the OMP, this algorithm can test the candidate solution got by iteration synchronously, which can avoid the error accumulation and improve the imaging quality.
Keywords: sparse data; range cell curvature correction; tomography; compressive sensing
层析SAR三维成像因其能有效解决传统SAR成像中有高度目标在距离?方位平面成像的模糊问题[1?2],而在如隐蔽目标探测、地表植被结构反演[3]等领域表现出了巨大的应用潜力。然而雷达的成像分辨率受到了奈奎斯特采样定理限制,这就要求无论在距离向、方位向的数据采集率还是在高度向上的阵元排布都要做到均匀而紧密,这会给A/D采样、雷达构型提出极大的挑战[4]。压缩感知理论提出只要信号具有稀疏性,就可以利用远少于奈奎斯特采样率的非均匀稀疏数据来高概率实现信号的无失真重建[5],该理论为降采样条件下的SAR三维成像实现提供了理论支撑,受到国内外雷达研究工作者的广泛关注[6?7]。……
