20

2026

-

07

Terahertz Compressed Imaging Based on Liquid-Crystal Programmable Metasurfaces

Author:


The team of Jingbo Wu and Xuemei Hu at Nanjing University has proposed a terahertz compressive imaging framework based on liquid-crystal programmable metasurfaces. This framework enables direct binary phase modulation, thereby implementing native symbol-mask encoding. Such phase-domain manipulation obviates the need for complementary measurements, significantly enhancing acquisition efficiency. To address the inherent symbol ambiguity associated with amplitude-only detection, the authors introduce a physics‑guided deep neural network for robust image reconstruction. Both simulation and experimental results demonstrate that high‑fidelity imaging can be achieved at sampling rates as low as 37.5%, offering a viable pathway toward efficient and scalable terahertz computational imaging.

The research findings were published on July 16, 2026, in Laser & Photonics Reviews, under the title “Phase-Modulated Terahertz Compressive Imaging With Physics-Integrated Deep Unfolding.”

Figure 1: Schematic diagram of phase-modulated terahertz compressive imaging based on a programmable metasurface and the PM-DUN framework.

Figure 2: Numerical comparison of compressed imaging performance under amplitude modulation and phase modulation.

Figure 3: The PM-DUN framework and its performance in image reconstruction.

Figure 4: Characterization of a liquid-crystal-based phase-modulating programmable metasurface.

Figure 5: PM-DUN framework for compressive sensing imaging

Source: Optics World