Azimuth Multichannel High-resolution And Wide-swath SAR Imaging Processing Technique

Synthetic aperture radar(SAR)is an imaging system which utilizes the relative motion between the radar platform and observation to achieve a fine azimuth resolution.The SAR system has a unique imaging capability and can work all-weather and all-day.With the development of remote sensing applications,SAR with highresolution and wide-swath(HRWS)imaging performance has become the development trend and research focus.However,the high azimuth resolution and wide swath are inherently contradictory.According to the Nyquist sampling theory,a fine azimuth resolution requires high pulse repetition(PRF).In comparison,wide swath requires a low operational PRF to avoid range ambiguities.Azimuth multichannel SAR system can be employed to resolve this contradiction and obtain HRWS SAR imaging.Generally,multiple uniformly spaced receiver channels are configured in multichannel SAR system.The single transmitter emits signal with low PRF to illuminate wide coverage area.All of the receiver channels simultaneously record the return signal.The lack of temporal sampling can be compensated by the additional spatial sampling and the effective azimuth sampling rate increases.As a consequence,we can improve the azimuth resolution,or increase the imaging swath without rang ambiguities,which means the realization of HRWS imaging.The echo data of each channel is aliased because of the low operational PRF and hence the reconstruction of unambiguous Doppler spectrum is necessary for HRWS SAR imaging.In real applications,however,the existence of channel errors deteriorates the performance of azimuth ambiguity suppression.Thus,the channel errors must be calibrated before azimuth reconstruction.In view of the increasing demand for high-resolution and wide-swath imaging of SAR systems,this paper focuses on the research of azimuth multichannel HRWS SAR imaging processing technology.This paper summarizes the imaging processing foundation of multichannel SAR.The main innovative work of this paper includes:Firstly,the fast channel error calibration algorithm is proposed.Based on the symmetric distribution of the azimuth antenna pattern of the SAR system,the covariance matrix of multichannel SAR signal in range-Doppler domain is a real Toeplitz matrix when there is no channel error and the Doppler centroid is compensated.With the real characteristic of covariance matrix,we can calculate the channel phase errors directly from the covariance matrix of real multichannel SAR signal.The proposed algorithm has the properties of high robustness and low computation load.Simulations and real data processing validates the effectiveness and superiorities of the proposed method.Second,a robust channel error calibration algorithm by maximizing the output power of MVDR beamformer is proposed.The essence of azimuth signal reconstruction can be considered as source separation,and MVDR beamformer can be used for azimuth signal reconstruction.When the steering vector is accurate,the output power of the MVDR beamformer is maximized.Based on this criterion,we construct a optimization problem and hence obtain the channel errors by using the inverse of covariance matrix and the nominal steering vectors.Compared to the subspace-based methods,the advantage of the proposed algorithm is that it does not require redundant channels and reduces system complexity;it avoids subspace swap phenomena and improves the robustness of the algorithm;it does not require covariance matrix decomposition and reduces the computation load.The performance of the proposed method is verified by simulations and real data processing.Third,based on the recurrent non-uniform sampling,an azimuth reconstruction algorithm is proposed.In the multi-channel SAR,the echo signal of each channel is uniformly sampled in azimuth,and the azimuth sampling of different channels during the same pulse may be nonuniform.Therefore,the azimuth sampling of multichannel SAR can be regarded as recurrent nonuniform sampling.Based on the recurrent non-uniformity of azimuth sampling and the theory of signal sampling,we derive the analytical expression of reconstruction filter coefficients.Compared with the DBF reconstruction methods,the proposed algorithm can directly obtain the reconstruction filter coefficients from the analytical expressions without matrix inversion.Thus,the algorithm reduces computational load.Simulations and measured data processing verify the effectiveness of the proposed algorithm.Fourth,the adaptive reconstruction method for azimuth signal is proposed.The azimuth sampling can be divided into three categories: uniform sampling,undersampling and over-sampling.The aliasing number can be calculated by azimuth crosscorrelation technique,and the effective system parameter of PRF,platform velocity and channel spacing can be estimated by using spatial spectrum estimation methods.With the aliasing number and effective system parameter,the ambiguity indexes of each Doppler bin can be determined and the azimuth signal can be adaptively reconstructed.Then the azimuth signal of multichannel SAR can be adaptively reconstructed.The adaptive reconstruction method can effectively suppress azimuth ambiguities even when the azimuth sampling is highly nonunifrom.Simulations and real airborne multichannel SAR data processing verify and analyse the effectiveness,superiorities and performance of the proposed method.