Novel SAR Autofocus Method Based on Deep Learning with GAN Network

This paper focuses on SAR autofocus based on deep learning. Firstly, considering the influence of phase on radar imaging results, aiming at the problem of defocus of SAR imaging results and poor focusing effect of some SAR imaging results by traditional autofocus methods, a SAR autofocus complex generation adversarial network model is constructed by using the method of deep learning, and the key parameters of the network are given. Then, the network training set and test set are constructed, and using the realistic measured SAR imaging results to train and test the network. Finally, the autofocus results of the network are displayed, and the autofocus results of GAN network are evaluated by using image entropy as an objective evaluation index, which proves the feasibility of using deep learning to solve the problem of SAR autofocus.