Ground Moving Target Indication For Multichannel SAR Systems

Synthetic aperture radar(SAR)is one of the high-resolution imaging radar systems and is widely used in civilian and military fields.Ground moving target indication has already been an important issue in synthetic aperture radar imaging.Normally,the key to detect moving targets is how to suppress stationary clutter clearly.With more freedom in the spatial domain,multichannel SAR systems can achieve better performance for clutter suppression and thus have been widely adopted to detect moving targets all over the world.The antenna structure and work mode of the multichannel SAR systems are exploited in detail in this paper.Besides,the difference of the focused location and phase of the stationary targets between multiple channels caused by displaced antenna phase center is analyzed in detail and an adaptive image registration algorithm is proposed subsequently.For multichannel SAR systems,moving targets are focused based on two different mode,which is related with their azimuth velocity(related with the Doppler rate).The moving target signal model can be used to analyze the distribution characteristics of the moving targets and to develop corresponding detection algorithms.A novel detection scheme is proposed for dual-channel SAR systems via exploiting the interferogram's magnitude and phase.The conventional ATI technique only considers the distribution character of the interferometric phase,and the detection performance suffers from noise.In this paper,the statistical distribution of the interferometric phase is exploited in detail and the relationship function between the interferometric magnitude and phase is presented based on the calculus.Two conclusions can be summarized : 1)interferometric phase is subject to Gaussian distribution and the fluctuations(variance)of the interferometric phase are inversely to the interferometric magnitude;2)the fluctuations of interferometric phase corresponding to the same interferometric magnitude are identical.Potential moving targets are indicated in the magnitude-phase plane with the developed 2-D detection threshold.The proposed algorithm can improve the detection performance of the conventional ATI technique and can achieve better detection performance of slow moving targets under the same SNR.Two detection algorithms are proposed for multichannel SAR systems based on spatial spectral analysis.Ideally,the spatial spectrum for multichannel SAR systems is a single frequency signal.The spatial spectrum of the stationary scatterer is distributed at zero,while that of the moving scatterer deviates from zero due to the radial velocity.In this paper,a new concept of dominant-velocity is proposed,which represents the motion properties of the dominant targets in each pixel.Moving targets are detected through statistics analysis of the dominant-velocity image(DVI).The mean? and the standard deviation ? are extracted from the DVI to describe the degree of dominant velocity consistency.Moving targets and stationary background are located at different region in the(?,?)plane and a parametric approach based on curve-fitting is used to separate moving targets from stationary backgrounds.However,this detection approach maybe fails to detect weak moving targets.A novel detection algorithm is proposed through measuring the asymmetry of the spatial spectra.It can be concluded that the spatial spectrum is symmetric around zero for a stationary scatterer but asymmetric around zero for a moving scatterer.A cost function is proposed to measure the asymmetry degree of the spatial spectra and the moving targets are indicated through threshold detection.In addition,the asymmetry of the spatial spectra can be measured directly in the spatial domain based on the Parseval's theorem and the procedure of estimating the spatial spectra can be avoided.Therefore,this algorithm is computationally efficient.Two detection algorithms suitable for high-resolution wide-swath SAR systems are proposed in this paper.High-resolution wide-swath SAR systems can achieve high resolution in azimuth and wide swath in range at the same time.The requirement analysis of moving targets detection for HRWS SAR systems is discussed in detail in this paper and the moving target signal model is also provided.Actually,the antenna structure of the HRWS SAR systems is the same with that of the multichannel SAR systems and conventional detection algorithms can be modified for HRWS SAR systems.Firstly,the two-look processing technique is applied to generate two sub-images and moving targets are detected through comparison of the obtained two sub-images.However,this algorithm is carried out without azimuth reconstruction for moving targets.Then,another detection algorithm is proposed and the conventional steering vector is generalized for moving targets through taking account the additional Doppler centroid shift caused by the radial velocity.Two modified steering vectors with symmetric velocity information are utilized to produce two images for the same scene.Due to the unmatched steering vectors,the stationary backgrounds are defocused but they still hold the same intensities in both images but moving targets are blurred to different extents.Therefore,ground moving target indication can be carried out via intensity comparison between the two images.