Intelligent Microwave Staring Correlated Imaging

Microwave staring correlated imaging (MSCI) is a super-resolution imaging technique based on temporal-spatial stochastic radiation fields (TSSRFs), which requires an accurate calculation of the electromagnetic field at the imaging plane. However, systematic errors always exist in practice, such as the time synchronization and frequency synchronization errors of radar systems, which make it difficult to calculate the required TSSRFs accurately, and this deteriorates the imaging results. Meanwhile, some imaging algorithms have problems such as high computational complexity. In this paper, an intelligent MSCI method based on the deep neural network (DNN) is proposed, which can accomplish imaging directly from the echoes, avoiding the computation of TSSRFs. A multi-level residual convolutional neural network (MRCNN) is developed for the DNN, and simulations and experiments are carried out to obtain the dataset for training and testing the MRCNN. Compared with the conventional MSCI methods, the imaging results verify the effectiveness of intelligent MSCI in terms of imaging quality and computational efficiency.