Research On Forward-looking Super-resolution Model And Imaging Method Of Airborne Radar

Airborne radar forward-looking imaging has important research significance and application value in the fields of formation target resolution,forward-looking terrain mapping and aircraft avoidance navigation.Limited by the physical size of the aircraft,the real aperture antenna of the airborne radar is small,the beamwidth is broad,and the angular resolution is poor,so the effective target information in the forward-looking area of the airborne platform cannot be accurately observed at a long distance.Due to the limitation of imaging mechanism,the synthetic aperture radar(SAR)and Doppler beam sharpening(DBS)techniques have imaging blindness in the forward-looking region.Airborne radar forward-looking super-resolution imaging technology can make use of the relationship between antenna pattern and target scattering contained in the scan beam echo sequences,and adopts signal processing method to obtain the target resolution beyond the real beam,which is the forefront of airborne radar forward-imaging.In order to improve the performance of the existing airborne radar forward-looking super-resolution imaging technology,it is urgent to make further breakthroughs in echo model,target adaptation and processing efficiency.In this dissertation,focusing on the the key issues of airborne radar forward-looking superresolution imaging technology,research has been carried out in echo modeling,target adaptation and efficient processing,and the main innovations are as follows:1.In this dissertation,a vector modulated convolution-like model is established.Based on the traditional amplitude convolution model,by introducing the Doppler phase matrix and the adaptive amplitude correction of the antenna pattern matrix,the convolution-like relationship between antenna pattern and target scattering is derived,and the vector modulated rules of azimuth echo sequences of moving platform radar are revealed.The proposed model reduces the model representation error and extends the velocity application range of echo model.Compared with the amplitude convolution model,the mean square error of the proposed model can be reduced by more than 3.5 d B,which provides basis for the research of super-resolution imaging methods.2.In this dissertation,a sparse target superresolution method based on singular value correction is proposed.Through the singular value truncation of the system measurement matrix and the non-convex norm regularization reweighted method,the sensitivity of the super-resolution inversion process to noise is reduced,and the super-resolution reconstruction of the sparse target under the condition of low signal-to-noise ratio is realized.3.In this dissertation,an extended target superresolution method based on generalized hybrid regularization is presented.Through the generalized hybrid regularization constraint term and adaptive iterative reweighted solver,the target scale reconstruction error is better than 1/10 beamwidth,and the accurate perception of the angle domain scale information of the extended target in the forward-looking area is realized.4.In this dissertation,a superresolution processing method based on matrix dimensionality reduction and matrix fast inversion is offered.Through the low-rank approximation matrix reduction and quick inversion of matrix identity,the complexity of superresolution processing is reduced fromO(N~3)toO(N~2),and the imaging processing time is reduced by more than 15 times,which meets the application requirements of forward looking super-resolution imaging for helicopter flight platform.The simulations and measured data have confirmed the effectiveness of the aforementioned echo model and methods,which can lay the theoretical and technical foundation for the engineering application of super-resolution imaging technology in terms of echo model,super-resolution methods and fast processing.