Wavenumber Spectrum Reconstruction Method for Microwave Computational Imaging With Re-Programmable Metasurface

Jinyu Zhang,Taiyang Hu,Xiaolang Shao,Hong Gu,Zelong Xiao
DOI: https://doi.org/10.1109/tci.2023.3265212
IF: 5.4
2023-04-21
IEEE Transactions on Computational Imaging
Abstract:Microwave computational imaging (MCI) employs uncorrelated illumination patterns via meta-materials to multiplex the information content of an entire scene into a few simple measurements, enabling image reconstruction through computational strategies. However, the reconstruction process often involves solving an ill-conditioned equation iteratively, which can result in a high computational burden and challenges in ensuring real-time performance. This paper presents a wavenumber spectrum reconstruction (WSR) method for fast imaging with reprogrammable metasurface (RMS). By constructing a series of code configurations, the electric field spectrum at the position of the RMS is measured. Then the wavenumber spectrum of the target is calculated by compensation in the wavenumber domain. Finally, the reconstruction of the target image is performed by inverse Fourier transformation (IFT) operation, which is less computationally burdensome and less time consuming. The proposed approach enables the construction of code configurations independent of the target's location and shape, thereby reducing the difficulty of optimizing measurement patterns. Simulation results show that the WSR method outperforms the traditional compressed sensing method in reconstructing target images with a low signal-to-noise ratio while significantly reducing computation time. The proposed method can satisfy the requirements for real-time performance and noise immunity in MCI systems, and can also be implemented on hardware platforms with limited computing resources.
engineering, electrical & electronic,imaging science & photographic technology
What problem does this paper attempt to address?