Study On Key Techniques For Circular SAR-GMTI

Circular Synthetic Aperture Radar(CSAR)is one of the special modes of Synthetic Aperture Radar(SAR).The aperture is formed by circular motion of the radar platform in CSAR.It has the capability of observation from full azimuth angle,sub-wavelength level high resolution and three-dimensional imaging.CSAR can provide long-term reconnaissance of the observation area,especially for the ground moving target in the observation area,it can track the ground moving targets for long time and obtain the trajectory information of them.The CSAR system with the above-mentioned Ground Moving Target Indication(GMTI)function is abbreviated as CSAR-GMTI.CSAR-GMTI has the functions of simultaneous stationary target imaging and GMTI.Main research contents of CSAR-GMTI include imaging,ground moving target detection,ground moving target trajectory reconstruction and ground moving target refocusing.This thesis has presented an in-depth research on the above issues,the main works and the results achieved in this thesis are as follows.In terms of CSAR system characteristics,the unique ring-shaped grating-lobes in CSAR is pointed out for the first time as much as the author's knowledge and its characteristics such as radius,width and magnitude are analyzed in detail.Furthermore,the relationship between Pulse Repetition Frequency(PRF)and ring-shaped grating-lobes is analyzed in theory,which has guiding significance for the design of CSAR system parameters.In the aspect of CSAR imaging,the Polar Format Algorithm(PFA)imaging method is introduced,and the two-dimensional auto-focus algorithm is involved into CSAR imaging,which effectively overcomes the problem of envelope errors which are neglected in traditional auto-focus algorithms.The slow ground moving target will leave a shadow in the SAR image,and the shadow detection can indirectly indicate the ground slow moving target trajectory.In this thesis,a ground moving target shadow detection method based on super pixel segmentation is proposed.This method can preserve the shadow structure information which is beneficial to shadow detection.Then,the detection results of multiple sub-apertures are clustered to provide the moving target trajectories.By discriminating the trajectories,false target trajectories are filtered out to achieve the purpose of false alarm suppression.The shadow of the fast ground moving target in the SAR image is weak or even absent,so the the reconstruction of the fast ground moving target trajectory is difficult to be obtained by aforementioned shadow detection based method.Aiming at this problem,this thesis proposes a moving target trajectory reconstruction algorithm based on dynamic programming with road grid constraints for single-channel CSAR system.Firstly,the fast moving targets are detected by Doppler filtering.Then,the trajectories in Range-Doppler domain can be obtained by tracking algorithms.Lastly,the dynamic programming is used to project the trajectories in Range-Doppler domain into road grid to achieve trajectory reconstruction.At present,the algorithm still has certain limitations,such as relying on prior road information and having no ability of the real-time processing,which will be addressed in the further works to improve the performance of the algorithm.In terms of ground moving target refocusing,a refocusing method based on parameter estimation for CSAR-GMTI is firstly proposed which is motivated by the usual methods in traditional linear SAR-GMTI field.Subsequently,the Inverse SAR technology is introduced into the CSAR-GMTI ground moving target imaging.By using the coherence of the CSAR system and the strong SNR of the ground moving target,a ground moving target imaging method based on frequency domain phase extraction was proposed.In addition,the use of Legendre Based on the of polynomial,an improved method based on Legendre polynomial fitting is also proposed.Due to the orthogonality of the Legendre polynomial,shortcomings of the proposed phase extraction method that requires pre-assumed polynomial order can be overcame.The algorithms proposed in this thesis are all verified by real data,including L,Ku,Ka,and THz band airborne SAR data,which proves the correctness and effectiveness of the algorithms.