Ground Moving Targets Detection and Parameter Estimation for Multi-channel SAR systems
Shengqi Zhu,G. Liao,Y. Qu,Zhengguang Zhou
2008-01-01
Abstract:Ground moving targets detection, imaging, positioning and motion parameter estimation in SAR has attracted the attention of many researchers all over the world. The author introduces a robust moving targets detection and parameter estimation approach for multi-channel SAR system which requires three consecutive operations in the following. Firstly, array calibration has been done among spatial channels. For multi-channel SAR system, the antenna patterns (amplitude and phase) in azimuth and elevation have different characteristic which degrades the correlation among spatial channels, leading to bad clutter rejection and inaccurate parameter estimation. Before data acquisition, T/R modules should be balanced using tested signal in order to ensure that point spreading functions (pdfs) are equivalent. Then transform the received echo data into azimuth frequency domain after range compression. Since array error varies with range, channel equalization should be done for every azimuth frequency and range segmentation divided predetermined. Eigenvector techniques or adjacent channel pair processing can be used [1]. Secondly, clutter rejection algorithm can be used to suppress clutter for detecting dim moving targets. Space Time Adaptive Processing (STAP) can be utilized to cancel stationary clutter. Post-Doppler reduced dimensional STAP algorithms such as Extend factored algorithm (EFA) [2] have good performance with small size of supported samples. CFAR detection can be used to residual clutter. Moving target occupies a straight line approximately in azimuth time domain. Thirdly, we propose an approach to estimate velocity and acceleration of target unambiguously. Conventional ML algorithm works well in across-track velocity estimation [1], across-track velocity also can be obtained directly from the interferometric phase between two images after clutter suppression [3]. Along-track velocity can be determined by the relationship between chirp rate and motion parameter. Time-frequency analysis tools, such as Wigner-Ville distribution (WVD) [4], Fractional Fourier transform (FrFt) [5], can be used to estimate Doppler rate. In order to avoid the cross-terms of the WVD in the presence of multiple moving targets, pseudo-WVD and smoothed pseudo-WVD combined with Radon transform have been proposed. However, all above algorithms have ignored the acceleration and the ambiguous phenomenon caused by across-track velocity and acceleration. The effect of along-track acceleration can be ignored since it exists in cubic or higher order terms, while the effect of across-track acceleration can not be ignored and influences the quadratic terms drastically. J.J.Sharma has analyzed the effects of target acceleration in dual-channel SAR-GMTI [6]. In this paper, we propose an approach to estimate