Research On Techniques Of Terrain Clutter Suppression And Moving Target Detection For Distributed Satellite Radar

Ground moving target indication(GMTI)or maritime moving target indication(MMTI)is one of the main tasks for airborne or spaceborne surveillance system in order to acquire the dynamic information from monitoring area.Combining synthetic aperture radar(SAR)technique and GMTI technique,surveillance system can obtain the high-resolution SAR image of stationary scene in the monitoring area as well as hold the dynamic information,thus to form the cognition and evaluation of the situation in the monitoring area,which has the important significance and worth.Compared with the single-channel system,the multichannel SAR system can integrate the spatial and temproal information source to improve the detection performance of moving target.Therefore,the multichannel SAR system become the main technique to achieve GMTI function.To further upgrade GMTI performance that is limited to the antenna aperture length due to the spaceborne single-platform system,the distributed satellite radar system becomes a good choice.Because of the existing of the long cross-track baseline(CTB),the distributed satellite radar is more sensitive to the undulating topographic elevation.Under the hybrid baseline condition,the undulating topographic elevation will result in the change of the structure of clutter spectra(i.e.,interferometric phase)in space position,in other words,the structure of clutter spectra in space position are contaminated since they are not only dependent-range(non-stationary)and dependent-elevation.The traditional processing methods are based on the homogenous environment and need the enough training(or secondary)samples that satisfy the independent and identically distributed(IID)conditions.However,the damaged IID environment will lead to the degradation of GMTI performance obviously.As a consequence,to deal with the problems of suppressing complicated terrain clutter and detecting moving targets for distributed satellite radar with the hybrid baseline in this dissertation,we focus mainly on these problems,including the selection of IID samples in the complicated terrain clutter region,the problem of performance deterioration of clutter cancellation resulting from the lack of samples,and the improvement of detection performance of moving targets with low signal-to-noise ratio(SNR).The main works can be summarized as follows in this dissertation:1.To address the selection problem of IID secondary samples(or the identification problem of azimuth ambiguities)for azimuth clutter suppression in the inshore region(i.e.,the boundary region between land and water),we propose a multi-feature(joint feature of multilook interferometric phase with magnitude)auto-segmentation-based approach is developed with the azimuth multichannel SAR systems.Firstly,the finite mixture clutter model for multilook covariance matrix(MLCM)is built,where the off-diagonal elements of MLCM contain the information of amplitude and interferogram's phase between azimuth channels.Then SAR image auto-segmentation is carried out by using the expectation maximum(EM)algorithm.Finally,azimuth ambiguous clutter can be suppressed by means of the clutter covariance matrix,which is constructed by the training samples of segmented ambiguities.The experiments on simulated data and real data measured by Terra SAR-X demonstrate that the proposed approach can obtain the more accurate position information and good cancellation performance for the azimuth ambiguous clutter,without the accurate system parameters and the information of the sources account for azimuth ambiguities.2.To deal with the lack of valid training samples after obtaining the IID samples in the complicated terrain clutter region,a generalized(refined)sample weighting approach in angle-Doppler domain(GSWADD)is presented for space-time adaptive processing(STAP),which is proved to be a generalization of sample weighting algorithms from the CCM reconstruction's perspective in angle-Doppler domain,to improve utilization ratio of sample data and effectively alleviate the STAP performance degradation of conventional weighting sample covariance matrix(WSCM)-type method.The core problem of proposed method is to get the clutter angle-Doppler spectrum by means of the local(refined)weighting coefficients and to reconstruct clutter-plus-noise covariance matrix(CCM)without aperture loss.Finally,the results of the experiments based on simulated and real measured data from airborne four-channel radar demonstrate that the proposed method can effectively alleviate the STAP performance degradation resulting from the loss of training samples contaminated by the interference targets,compared with the conventional statistical STAP algorithms.3.To address the problem of the degradation of detection performance for a moving target with a low SNR for distributed satellite radar systems,we develop a SAR-GMTI approach(or framework)based on the aid of moving-target shadow feature with multichannel and high-resolution SAR systems.First of all,the models of a moving-target shadow in dimensions,intensity,and interferometric phase are introduced.And then we propose the two kinds of methods to improve the detection performance of moving-target shadow,on the basis of the presented fundamental principle and implementation framework of the shadow-aided decision with multichannel SAR systems.The first method is proposed to improve the detection of shadow according to the fusion of amplitude and interferometric phase features.And the second one is developed to ameliorate the shadow detection performance by exploiting a priori information including moving-target velocities and dimensions to adaptively adjust thresholds in the moving-target shadow region,and hence to improve the detection performance of the proposed shadow-aided decision method.Finally,numerical experiments on simulated and real data by airborne dual-channel radar are presented to demonstrate that the developed framework can obtain good results for the detection of the extended ground moving target with low SNR.Note that the proposed method is implemented by firstly detecting moving targets with low threshold(both the probabilities of false alarm and detection will meanwhile rise),and then by eliminating the isolated false alarms according to the shadow feature of the moving target.