Nonrandom Microwave Ghost Imaging

Ghost imaging (GI) is a novel imaging method originated from quantum and optical areas. Due to its nonlocal and nonscanning features, the GI has drawn much attention from different research communities and been adopted into microwave imaging scenarios recently. However, most of the current microwave GI schemes reported in the literature are mainly theoretical approaches. On the one hand, this is because conventional microwave GI systems require a large number of antennas transmitting random modulated signals which is quite challenging in practice. On the other hand, unlike optical GI where charge-coupled devices are used to collect the spatial information of background light fields, current microwave GI schemes can only rely on the estimation of the background microwave fields. Consequently, the reconstruction of objects will suffer from extra perturbations due to estimation errors. Therefore, in order to reduce the complexity and difficulty in the practical implementation of microwave GI, we proposed a novel microwave GI scheme based on nonrandom electromagnetic (EM) fields in this paper. By applying purposely designed EM fields to illuminate the imaging scenario, both the requirement of randomness and the involvement of background microwave field estimations in the framework of current microwave GI have been removed. Thus, its mathematical imaging model has been simplified from a totalleast-square (LS) problem to an ordinary-LS one. In addition, by the employing binary orthogonal matrix as the reference for generating the background EM fields, the reconstructed image can be obtained directly without iterative refinement. Numerical simulations show that the proposed nonrandom microwave GI can effectively reconstruct objects with different profiles under different signal-to-noise-ratio conditions. It also shows that both the reconstruction performance and complexity of the proposed method are superior to conventional microwave GI. Besides, similar comparison results can be observed when compressive samplings are applied.