Research On Algorithms Of Motion Compensation And Imaging For Sparse Aperture ISAR

Inverse synthetic aperture radar(ISAR)plays an increasingly important role in military and civil fields,such as strategic defense,space detection,and port surveillance,since it can perform long-distance imaging of targets under all-weather and day-and-night conditions.With the increasing demands for these applications,high resolution ISAR imaging with high efficiency is of great urgency.To this end,conventional motion compensation and imaging techniques should be improved and updated.In practical applications,due to the influence of many external factors including fierce space countermeasures and complex interference of electromagnetic environment,the echo data received by the radar may encounter pollution and loss.In addition,the multi-task cooperative working mode of the radar system or other reasons can cause the loss of azimuth aperture in the echo data.Under the conditions of sparse aperture(SA),traditional imaging algorithms often suffer from low resolution and severe sidelobe interference.Furthermore,SA is highly correlated with the accuracy of motion compensation and cross-range scaling.Therefore,it is of great practical significance to study algorithms with high imaging efficiency and quality under the SA conditions.In this thesis,a systematic study of motion compensation and imaging under the SA conditions is carried out.The imaging part contains target imaging and cross-range scaling of the ISAR images,aiming to improve the imaging efficiency and quality.The main contents include:(1)fast phase adjustment,(2)superresolution imaging of steady targets,(3)high-resolution imaging of maneuvering targets,and(4)cross-range scaling algorithm for the scaling of slow rotation targets.The summaries of the projects are as follows:(1)Motion compensation algorithms suitable for SA data are given,and a fast SA-ISAR phase adjustment algorithm based on eigenvector is proposed to avoid the high computational complexity of the conventional eigenvector-based phase adjustment algorithm.The eigenvector corresponding to the maximum eigenvalue of the data covariance matrix after range alignment already contains the information for phase adjustment.While the original phase adjustment algorithm usually requires a large amount of computation,since it directly performs eigen-decomposition on the high-dimensional data covariance matrix.The proposed fast algorithm utilizes Lanczos or Arnoldi iteration to solve the target eigenvector,which avoids the high computational eigen-decomposition step,thereby greatly reducing the computational complexity and improving the efficiency.Experimental results indicate that the proposed algorithm has a lower computational complexity while ensuring the accuracy of phase adjustment.(2)Two SA-ISAR super-resolution imaging algorithms,including SA-ISAR imaging algorithm based on the multiple signal classification(MUSIC)and SA-ISAR imaging algorithm based on the spectral-factorization root-MUSIC(SF-Root-MUSIC),for steady targets are developed,aiming to solve the issues of low resolution of the conventional imaging algorithms and poor imaging quality under sparse conditions.The MUSIC-based SA-ISAR super-resolution imaging algorithm first restores the data covariance matrix of each range bin to a full-rank matrix through spatial smoothing.Then,the scatterer number is estimated by the Gerschgorin disk estimator,while the positions of the scatterers are evaluated by the MUSIC algorithm.Finally,super-resolution imaging results are obtained by calculating the intensities of the scatterers in light of the least squares algorithm.The SF-Root-MUSIC-based SA-ISAR super-resolution imaging algorithm employs polynomial rooting to estimate the positions of the scatterers,where polynomial rooting is used to replace the spectrum search process of the MUSIC algorithm.In the SF-Root-MUSIC-based algorithm,the order of the polynomial is reduced by half through the spectral-factorization,thereby significantly decreasing the computational complexity and improving the imaging efficiency.Experimental results demonstrate that the proposed algorithms possess high imaging reso lution and can effectively improve the imaging quality under the SA conditions.(3)A SA-ISAR high-resolution imaging algorithm based on the Relaxation method is proposed,which is aiming at the imaging of maneuvering targets.According to the turntable model of the maneuvering targets,the echo signal of each range bin can be regarded as a multi-component linear frequency modulated(LFM)signal after motion compensation.The proposed algorithm estimates the parameters of the multi-component LFM signal using the Relaxation method,and then,the range instantaneous Doppler imaging results are obtained by analyzing the instantaneous Doppler frequency of the echoes from the scatterers at different time intervals.The proposed algorithm has high imaging quality and strong robustness to the data missing.When a large amount of data is missing,it can still give well-focused imaging results.Experimental results indicate that the algorithm is highly effective.(4)A MUSIC-based SA-ISAR cross-range scaling algorithm is developed to overcome the difficulty of cross-range scaling of slow rotation targets.The algorithm first chooses the range bins with single strong scatterer on the basis of the amplitude normalized variance criterion,and the echo signals corresponding to these range bins can be regarded as single-component LFM signals.Subsequently,the LFM signal of each range bin is reduced in order,and the frequency of the reduced signal is accurately estimated using the MUSIC algorithm.After processing all selected range bins,a linear relationship related to the target rotation velocity can be obtained.Eventually,the ISAR image can be cross-range scaled in terms of the target rotation velocity.The proposed algorithm can effectively achieve the cross-range scaling of slow rotation targets with high accuracy and low computational complexity.Furthermore,the algorithm can be used in the case of cross-range scaling for the SA data.Experimental results illustrate that this novel algorithm is advantageous over the conventional algorithms.