Study On Azimuth Multichannel SAR Image Domain Reconstruction And Moving Target Processing

In the history of the development of SAR,the pursuit of the two key indicators,that is resolution and swath,which ave always been the two most important directions.The improvement of these two indicators is of great significance to enhance the ability of SAR radars to rapidly acquire fine feature and global information of targets,and to realise military and civil applications such as military reconnaissance,homeland resource observation and disaster assessment over a wide range.Conventional SAR with single-transmission and single-receiver is limited by the minimum antenna area and cannot achieve both high resolution in azimuth and wide swath in range direction at the same time.Azimuthal multichannel SAR combined with digital beamforming technology is an effective way to slove this problem.However,due to the manufacturing process,inconsistency between multiple channels is an objective in real systems,and this inconsistency will eventually seriously affect the reconsrtuction and imaging quality.Traditional multichannel SAR processing methods are generally performed in the raw signal domain,i.e.before imaging.In this paper,a new idea of processing in the image domain is proposed.In comparison,image-domain multichannel processing has many advantages such as high local signal-to-noise ratio,more sensitive to error variations compared to the raw signal,and localisation of the scene echo signal.Therefore,this thesis proposes a new imagedomain approach to multi-channel SAR high-precision signal processing,focusing closely on the theory and methods of modelling and analysis of image domain signals,ambiguityfree image domain reconstruction,image domain high-precision channel error estimation and image domain moving target processing.In Chapter 2,the problem of modelling the image domain signal in the presence of the azimuth ambiguity is addressed.Based on the modelling of azimuthal multichannel SAR,the expressions in the image domain of the single channel echo signal are derived by using the classical Range Doppler algorithm.The approximate calculation ways of image domain information of ambiguous points relative to the true points,such as position offset,amplitude attenuation and shape in azimuth and range dimensions is given,which can be used to quantitatively determine the position or amplitude information of the ambiguous point which can provide a priori information for the detection and identification of ambigous targets in high-resolution wide-sawth SAR images;in addition,this chapter gives the basis and constraints for the classification of the shape of ambigous points,which can be used to determine the shape class of ambigous points under the condition that the system parameters are known.For the problem of multi-channel SAR high-precision phase error estimation,an image subspace estimation method is proposed in Chapter 3.Firstly,the mechanism by which phase error affects multi-channel signal reconstruction and ambiguity suppression is analysed to provide a basis for subsequent image-domain phase error estimation.The method obtains the multichannel SAR images for error estimation through the steps of zero-padding,reconstruction and imaging;then,after the image domain degree of freedom compression,the channel phase error is estimated using the statistical image subspace;finally,the multichannel images are error corrected.An ambiguity-free high-resolution wide-swath SAR image is obtained.In this chapter,we find the phenomenon that image domain degree of freedom doubles by using simulation data,and provide a conclusion that degree of freedom doubles in the image domain and the proposed degree of freedom compression method.The validity and advantages of the image-domain error estimation method proposed in this chapter are verified by processing and evaluating the simulated and measured data.To address the problem of computationally intensive image domain subspace methods,Chapter 4 proposes an algorithm for channel phase and baseline error estimation based on image subspace-least squares(ISP-LS).The method is completed with three main steps:firstly,pre-processing and SAR imaging;secondly,the ISP-LS-based channel phase and baseline error estimation and calibration algorithm;and thirdly,post-imaging reconstruction and ambiguity suppression.In order to reduce the increased computational complexity due to channel zeroing,the SAR imaging process in this chapter directly processes the original un-zero-padded echo data,and the jointly estimate the channel phase and baseline errors based on image subspace after SAR imaging,which also has the advantages of image domain processing,can use the part of the SAR image with higher signal-to-noise ratio to estimate the image subspace and achieve more accurate image domain error estimation with relatively lower computational effort.The improvement in computational complexity of the proposed method is verified by evaluating the computational load in this chapter.In addition,as a key step in the improved image domain high accuracy error estimation process,an image domain reconstruction method based on multi-channel SAR images is also proposed,which can be achieved by first imaging each channel data and then combining these multi-channel SAR images to achieve image domain reconstruction and ambiguity suppression.Finally,simulated as well as measured airborne multi-channel SAR data are used for processing and compared with the conventional ones to verify the effectiveness of the image-domain method and its advantages in terms of error estimation.Chapter 5 addresses the problem of inconsistent signal characteristics of moving targets and scenes in multichannel SAR images resulting in the inability to reconstruct them simultaneously.This chapter proposes a method for simultaneous reconstruction of moving targets and stationary scene targets in the multi-channel SAR image domain.The reconstructed multi-channel SAR image is unable to reconstruct the moving target without ambiguous the azimuthal spectrum,so the moving target is scattered in the final image and there are multiple ambiguous points along the azimuthal direction.Traditional space-time adaptive processing methods to reconstruct and image both moving targets and stationary scene signals simultaneously are often difficult to meet in practice in terms of spatial freedom.Since the moving target is partially focused and distributed in the SAR image,different steering vectors can be used at different ambiguous moving target positions,effectively reducing the need for spatial degrees of freedom.Finally,simulation experiments verify the effectiveness of this method.