Research On Non-Iterative Super-Resolution Imaging Algorithm Of Airborne Forward-Looking Scanning Radar

Airborne scanning radar can obtain images of the forward-looking area of the flight platform,and has important application value in the fields of autonomous landing,terrain avoidance,maritime search,and ground attack.However,the azimuth resolution of traditional real-beam radar images is often limited by the range and antenna size of the radar,and azimuth super-resolution imaging can be achieved through convolution inversion.At present,most of the super-resolution algorithms are iterative methods,which have problems such as complex update,slow convergence speed,long processing time,and difficult parameter setting.In view of the above problems,this thesis focuses on the non-iterative superresolution imaging algorithm of airborne forward-looking scanning radar.The research proposes a multi-dimensional filter wavelet threshold super-resolution imaging algorithm and a super-resolution imaging algorithm based on generative adversarial network.The main contents are as follows:1.According to the working mode of the airborne scanning radar,the echo signal characteristics of the airborne scanning radar are analyzed,and the convolution model of the radar antenna pattern and the target scattering coefficient is established.The impact of super-resolution algorithm performance provides theoretical guidance for follow-up work.2.The multi-wiener-wavelet threshold super-resolution imaging algorithm is proposed,the Wiener filtering and its improved algorithm are studied,the superresolution processing is transformed into a multi-parameter deconvolution unit weighted sum by using the linear expansion theory,and the minimum Stein Unbiased estimation solves the weight coefficients,reduces noise sensitivity through wavelet thresholding,and realizes adaptive deparameterized forward looking super-resolution imaging.At the same time,the algorithm has good robustness and scene adaptability.3.Proposed a super-resolution imaging algorithm based on generative adversarial network,built a generator network and discriminator network architecture for superresolution imaging,used VGGNet-19 for network pre-training,and designed the network loss function by introducing the antenna pattern.Forward-looking super-resolution imaging under low signal-to-noise ratio is achieved.The simulation results show that the algorithms proposed in this thesis can quickly realize the super-resolution imaging of airborne forward-looking scanning radar,and have the characteristics of noise self-adaptation and deparameterization.