Guidance image based method for real-time motion artefact handling on Time-of-Flight cameras

This paper presents a real-time approach to address the motion artefact related to the Time-of-Flight (ToF) camera's working principle. ToF cameras based on demodulation lock-in pixels estimate depth from the phase-shift between emitted and received modulated near-infrared (NIR) signals, in which four sequential phase-shifted images are required, i.e., four-taps technique. The ToF working principle assumes the scene to be motionless during this time interval. However and in practise, unreliable depth measurements arise along object boundaries in dynamic scenes, mainly when fast movements are involved. Herein, we propose a robust method to identify those pixels in the resulting depth map that are prominent to be unreliable. Then, we replace their values by the closest reliable ones using the guided filter (GF) and an accurate guidance image, generated from the previously acquired sequential phase-shifted images. The GF has been selected as it presents a better behaviour near edges than alternative edge preserving filters with a major advantage of being a fast and non-approximate linear time algorithm. The experimental evaluation shows that the proposed method satisfactory addresses the motion artefact, even in extreme conditions.