Making underwater imaging a little less murky

In cloudy, murky conditions — such as underwater environments — light signals are weak. The ability to use photons for communication or other applications requires accurate instruments and efficient processing. Zeng et al. developed a system to accomplish just that.

“Suspended particles, bubbles, organic matter, and other media will have a strong scattering and absorption effect on the optical signal, resulting in the deviation of the photon path,” said author Yaoxing Bian. These can lead to the signal being distorted, blurred, and reduced.

The system can clearly recover the phase information of a vortex beam at 0.022 photons per imaging pixel. In tests with turbidity ranging from 0.1 to 100 Nephelometric Turbidity Units (NTU) — a quantitative measurement of a liquid’s cloudiness — the researchers’ technique adapted to different levels of noise and varying environmental conditions.

The scientists developed a neural network to recover low-quality images, which they combined with existing single-pixel wavefront imaging techniques. To make their technique as general as possible, they trained their model on simulated photon measurements with the types of noise expected to be encountered underwater, like partial obscurement or random image cropping. They tested its application on clean tap water with low NTU, then added milk to achieve higher NTU.

Improved photon signals can be used everywhere from underwater communication to particle physics experiments.

“This study demonstrates the great potential released by the deep integration of artificial intelligence and optical imaging systems, and provides feasible solutions for underwater rescue, quantum sensing, and secure communication,” said Bian. “In the future, we will further expand the application capabilities of the system in dynamic image reconstruction and three-dimensional imaging and explore its potential value in a wider range of frontier fields.”

Source: “Photon-level single-pixel wavefront imaging through turbid underwater environment,” by Fanjin Zeng, Yaoxing Bian, Kai Song, Hongrui Liu, Shijun Zhao, Xingyu Wang, Futong Zhang, Hongda Ge, Dong Wang, and Liantuan Xiao, APL Photonics (2025). The article can be accessed at https://doi.org/10.1063/5.0273495 .

This paper is part of the Single-photon detectors: new physical principles, circuits and applications collection, learn more here .