Super-resolution Orthogonal Deterministic Imaging Technique for Terahertz Subwavelength Microscopy

TL;DR

This paper introduces SODI, a super-resolution imaging method for terahertz microscopy that uses designed masks to enhance resolution with minimal frames, inspired by super-resolution techniques in visible light.

Contribution

The paper develops a novel super-resolution technique for THz imaging using artificial masks, achieving higher resolution with fewer frames than existing methods.

Findings

Experimental second-order resolution enhancement with 8 frames

Theoretical extension to higher-order resolution improvements

Method adaptable to existing THz imaging systems

Abstract

Terahertz subwavelength imaging aims at developing THz microscopes able to resolve deeply subwavelength features. To improve the spatial resolution beyond the diffraction limit, a current trend is to use various subwavelength probes to convert the near-field to the far-field. These techniques, while offering significant gains in spatial resolution, inherently lack the ability to rapidly obtain THz images due to the necessity of slow pixel-by-pixel raster scan and often suffer from low light throughput. In parallel, in the visible spectral range, several super-resolution imaging techniques were developed that enhance the image resolution by recording and statistically correlating multiple images of an object backlit with light from stochastically blinking fluorophores. Inspired by this methodology, we develop a Super-resolution Orthogonal Deterministic Imaging (SODI) technique and apply it in the THz range. Since there are no natural THz fluorophores, we use optimally designed mask sets brought in proximity with the object as artificial fluorophores. Because we directly control the form of the masks, rather than relying on statistical averages, we aim at employing the smallest possible number of frames. After developing the theoretical basis of SODI, we demonstrate the second-order resolution improvement experimentally using phase masks and binary amplitude masks with only 8 frames. We then numerically show how to extend the SODI technique to higher orders to further improve the resolution. As our formulation is based on the equation of linear imaging and it uses spatial modulation of either the phase or the amplitude of the THz wave, our methodology can be readily adapted for the use with existing phase-sensitive single pixel imaging systems or any amplitude sensitive THz imaging arrays.