LMMSE 3-D SAR Focusing
LMMSE 3-D SAR Focusing
复制标题
DOI：
10.1109/tgrs.2008.2000633
复制
发表时间：
2009
影响因子：
8.2
通讯作者：
G. Fornaro;A. Pauciullo
中科院分区：
文献类型：
--
作者：
G. Fornaro;A. Pauciullo
三维合成孔径雷达(SAR)成像，也被称为合成孔径雷达层析成像，它利用多个视角，通过获取不同高度的散射功率分布，将合成孔径雷达系统的成像能力扩展到三维成像。目前的卫星技术是通过在同一场景上连续通过单天线SAR传感器来获得多个视图，但预计下一代传感器编队将获得多基地数据。波束形成或奇异值分解反演等常规处理是基于几何导数的，因此假设精确的相位校准和不存在目标去相关。本文分析了残差、未补偿大气贡献和时间去相关对相位误差的影响，提出了一种基于线性最小均方误差方法的三维成像技术。所得到的算法通过执行考虑先验数据相关性属性的数据集成来扩展传统处理的可能性。因此，它允许处理附加随机贡献的存在，例如：时间相干损失和大气相位错误校准。此外，参考未来的双基地和多基地系统，它允许对重复通过同时天线采集的数据进行改进的相干积分。
Three-dimensional synthetic aperture radar (SAR) imaging, a technique also known as SAR tomography, uses multiple views to extend the capability of SAR systems to 3-D imaging by achieving a profiling of the scattering power at different heights. Multiple views are obtained with the current satellite technology via successive passes of a single antenna SAR sensor over the same scene, but next-generation sensor formations are foreseen to acquire multistatic data. Conventional processing, such as the beamforming, or singular values decomposition inversion is based on geometrical derivations and, hence, assumes the accurate phase calibration and the absence of target decorrelation. This paper analyzes the effects of phase miscalibration due to residual uncompensated atmospheric contribution and temporal decorrelation and proposes a 3-D imaging technique based on a linear minimum mean square error approach. The resulting algorithm extends the possibilities of the conventional processing by carrying out an integration of data that accounts for the a priori data correlation properties. Hence, it allows handling of the presence of additional stochastic contributions such as: temporal coherence losses and atmospheric phase miscalibration. Moreover, with reference to future bistatic and multistatic systems, it permits an improved coherent integration of data acquired by simultaneous antenna in repeated passes.