Non-Line-Of-Sight (NLOS) imaging aims at recovering the 3D geometry of objects hidden from the direct line of sight. This paper introduces a multipixel time-of-flight NLOS imaging method combining Single Photon Avalanche Diode (SPAD) array detectors with a fast reconstruction algorithm. This allows for capturing and reconstructing live low-latency videos of NLOS scenes with natural, non-retroreflective objects. A signal-to-noise-ratio (SNR) model for NLOS imaging is developed and used to devise a reconstruction method where SNR, motion blur, angular resolution, and depth resolution are independent of scene depth.
Publisher
Nature Communications
Published On
Nov 11, 2021
Authors
Ji Hyun Nam, Eric Brandt, Sebastian Bauer, Xiaochun Liu, Marco Renna, Alberto Tosi, Eftychios Sifakis, Andreas Velten
Tags
Non-Line-Of-Sight
3D imaging
SPAD detectors
reconstruction algorithm
signal-to-noise ratio
low-latency video
natural scenes
Related Publications
Explore these studies to deepen your understanding of the subject.