Missile-Borne Forward-Looking Bistatic Synthetic Aperture Radar Imaging

Bistatic synthetic aperture radar(BSAR)imaging technology uses the relative motion between the transceiver and the target to form a high-resolution image.As a kind of microwave active imaging system,it has the characteristics of all-weather,all-time,high resolution,strong penetration.It has a wide range of applications in military and civil fields.Because the application scenario is mainly high mobility platform,this paper uses high mobility platform application scenario modeling,focusing on missile-borne forward-looking static target area imaging and missile-borne forward-looking moving target area imaging,and uses Convolution Neural Network(CNN)to optimize the image obtained by radar algorithm,so as to improve the resolution of radar image.The main contents of this paper are as follows:Firstly,the reason why missile-borne forward-looking imaging uses BSAR is introduced,and the bistatic slant range expression of the system is given from the perspective of imaging scene modeling.Then,the two-dimensional spectrum expression of echo signal is derived by using the Principle of Stationary Phase(POSP).In order to simplify the complexity of the imaging filter,the rationality of using the third-order slant Taylor expansion is verified.The rationality of the algorithm is verified by simulation.Secondly,in the process of missile guidance,the target needs to be located in real time.The detonator will look forward at the target area during the flight,and the imaging distance becomes closer as the missile approaches the target.The image quality of bistatic SAR imaging algorithm will deteriorate rapidly when the detonator enters the fast dive phase.The main reason is defocusing in azimuth direction,and the more defocusing from the center of imaging aperture,the more serious defocusing.In this paper,an improved CNN structure based on u-net is proposed,and the composition of the network structure is explained.Then the generation process and rationality of the training set are described,and the overall implementation steps of the experiment are listed.Finally,the effectiveness of the proposed method is verified by defocused point target and fuzzy ship surface target with additional noise.Finally,for the non-cooperative target(moving target),in the specific imaging,because the synthetic aperture radar needs a certain amount of time for azimuth slow time echo accumulation,which is usually second level,so the target’s motion characteristics will introduce the echo,resulting in the imaging can not compensate the additional phase caused by the motion in the echo,and the imaging quality is greatly reduced.In this paper,the idea of Fourier Ptychographic Microscopy(FPM)in optics is used for reference,and all the motion characteristics of non-cooperative targets are processed intensively,and then CNN is used to achieve focusing effect.