Research On Sparse-based MIMO-ISAR Three-dimensional Imaging Technology

The microwave imaging technology,including ISAR,SAR,In ISAR and In SAR,is formed with the combination of radar's inherent all-time all-weather working characteristics and modern signal processing theory represented by Fourier analysis.It is independent of climate and time,providing rich multi-dimensional features and structural information of the target.Therefore,it has important application prospect in many military and civilian fields.As a kind of new radar system,Multiple-input Multiple-output(MIMO)radar observes the target by multiple transmitting and multiple receiving antennas simultaneously.It can acquire a number of observation channels more than the actual antenna elements to form a spatial virtual aperture.So it is a research hotspot in the radar field in the past decade.MIMO-ISAR three-dimensional imaging technology owns the advantages of ISAR imaging and MIMO radar imaging.However,there are many unobserved positions inevitably in the equivalent global observation samples when jointing the time samples and space samples,which forms the observation sparse problem.In addition,the antenna array may also suffer from array sparsity caused by partial antennas failure,damage and so on in practical applications.Therefore,this paper focuses on two typical sparse problems above in MIMO-ISAR three-dimensional imaging technology.The specific research work is shown as follows:In view of the sparse observation problem,this paper starts from the perspective of multi-snapshot images joint utilization.Based on the MIMO radar single-snapshot three-dimensional imaging technology,the cosine information of the target's movement direction is estimated.And multiple single-snapshot three-dimensional images,obtained in a specific period of observation,are integrated along the estimating direction.An improved three-dimensional image is reconstructed by extracting all peak slices.Simulation experiments demonstrate that the method overcomes the sparse observation problem,is able to image the unmanned aerial vehicle(UAV)target effectively and improves the azimuth resolution of the movement direction.In view of the sparse antenna array problem,this paper considers the situation that the number of missing antennas increases from small to large.When few antennas are missing,the basic matrix completion method is used to recover the sparse observation samples.As the number of missing antennas grows(such as the entire row or column is missing),an extended matrix completion method is proposed,adding the priori information that the scattering point of the radar image is sparse.Simulation experiments show that different matrix completion methods can effectively recover observation samples for respective array conditions.And a clear image of the target can be obtained finally.