An Efficient Compressive Sensing-Based Method for Microwave Inverse Imaging Using Sparse Induced Current

By leveraging the relation between sampling and sparsity, compressive sensing (CS) has been developed and applied in many fields, including electromagnetics. In this article, we propose a CS-based method (CSM) to quantitatively image unknown objects under test (OUTs) within a real-time period. Throughout the resolve of inverse scattering problem (ISP), the CS scheme is used to estimate the sparse induced current and contrast profiles of the OUTs, respectively. An experimental system assisted with switching network was established to acquire the scattered data without any mechanical movement. The simulations and experiments verified that the proposed method is able to reconstruct the quantitative images when solving ISPs for sparse objects. In comparison, it is found that the constructed images by CSM are overall superior to the three typical traditional inverse methods, while providing a clean background with less artifacts. Based on CSM, the images of four testing experiments can be reconstructed in nearly 0.35 s with a minimum mean square error of 0.051, 0.085, 0.093, and 0.131, respectively. In terms of computational efficiency, the millisecond-level inversion of CSM implies promising potential in the application of real-time imaging.