Non-line-of-sight camera gets a resolution boost

To see what’s around a corner or other obstacle, periscopes depend on mirrors or prisms. Non-line-of-sight (NLOS) imaging relies instead on detecting single photons reflected off several surfaces and reconstructing the image algorithmically. The method is limited by the temporal resolution of single-photon detectors (SPDs), which is typically in the dozens of picoseconds. That time scale restricts the image resolution to about a centimeter at best.

But by using a short laser-pulse pumping technology, a group led by Jian-Wei Pan, Qiang Zhang, and Feihu Xu of the University of Science and Technology of China have increased SPD precision by an order of magnitude compared with previous results . The researchers demonstrated the improvement by reconstructing letters of a few millimeters in size, shown in the photo.

For the technique to work, the photons from the signal laser scatter first off a wall in the line of sight and then from a hidden surface parallel to the wall, shown in the diagram. A dichroic mirror filters the pump laser so that it couples to the signal laser’s path. From there, photons from the signal and pump lasers travel through a waveguide to a photodiode that detects the photons.

The researchers scanned the delay between the pump laser and signal laser at 0.2 ps increments to determine precisely when the individual photons hit the detector. That strategy yielded an average arrival-time precision of 1.4 ps and was critical to differentiating the photons scattered off the visible wall from those scattered by the hidden wall.

The researchers analyzed the method’s sensitivity to the detector’s time resolution and found that the letters written on the hidden wall were illegible at a time scale of 5 ps. The next-best NLOS imaging system uses a superconducting nanowire SPD and posted a resolution of about 4 ps but requires cryogenic cooling. The new imaging method operates at room temperature and could thus prove useful in various applications, including medical imaging and manufacturing. (B. Wang et al., Phys. Rev. Lett. 127, 053602, 2021 .)