This paper demonstrates mid-infrared (MIR) single-photon computational imaging using a single-element silicon detector. The method employs nonlinear structured detection with encoded time-varying pump patterns imprinted onto the MIR object image via sum-frequency generation. This translates the MIR radiation into the visible region, enabling single-photon upconversion detection. Compressed sensing and deep learning algorithms reconstruct MIR images under sub-Nyquist sampling and low-light conditions. The system achieves single-pixel simplicity, single-photon sensitivity, and room-temperature operation, opening new avenues for sensitive imaging at longer infrared wavelengths or terahertz frequencies.

Yinqi Wang, Kun Huang, Jianan Fang, Ming Yan, E Wu, Heping Zeng