Abstract

Collecting higher-quality three-dimensional points-cloud data in various scenarios practically and robustly has led to a strong demand for such dToF-based LiDAR systems with higher ambient noise rejection ability and limited optical power consumption, which is a sharp conflict. To alleviate such a clash, an idea of utilizing a strong ambient noise rejection ability of intensity and RGB images is proposed, based on which a lightweight CNN is newly, to the best of our knowledge, designed, achieving a state-of-the-art performance even with 90 × less inference time and 480 × fewer FLOPs. With such net deployed on edge devices, a complete AI-LiDAR system is presented, showing a 100 × fewer signal photon demand in simulation experiments when creating depth images of the same quality.

© 2023 Optica Publishing Group

228 × 304 200-mW lidar based on a single-point global-depth d-ToF sensor and RGB-guided super-resolution neural network

Miao Sun, Yifan Wu, Lebei Cui, Hengwei Yu, Jie Li, Jian Qian, Jier Wang, Lei Qiu, Patrick Yin Chiang, and Shenglong Zhuo
Opt. Lett. 48(13) 3415-3418 (2023)

SPADnet: deep RGB-SPAD sensor fusion assisted by monocular depth estimation

Zhanghao Sun, David B. Lindell, Olav Solgaard, and Gordon Wetzstein
Opt. Express 28(10) 14948-14962 (2020)

Non-fusion time-resolved depth image reconstruction using a highly efficient neural network architecture

Zhenya Zang, Dong Xiao, and David Day-Uei Li
Opt. Express 29(13) 19278-19291 (2021)

Monocular depth estimation network with single-pixel depth guidance

Hongjae Lee, Jinbum Park, Wooseok Jeong, and Seung-Won Jung
Opt. Lett. 48(3) 594-597 (2023)

High-definition image transmission through dynamically perturbed multimode fiber by a self-attention based neural network

Guohua Wu, Yong Sun, Longfei Yin, Zhixiong Song, and Wenting Yu
Opt. Lett. 48(10) 2764-2767 (2023)