Analyzing interference between RGB-D cameras for human motion tracking

Abstract

Multi-camera RGB-D systems are becoming popular as sensor setups in Computer Vision applications but they are prone to cause interference between them, compromising their accuracy. This paper extends previous works on the analysis of the noise introduced by interference with new and more realistic camera configurations and different brands of devices. As expected, the detected noise increases as distance and angle grows, becoming worse when interference is present. Finally, we evaluate the effectiveness of the proposed solutions of using DC vibration motors to mitigate them. The results of this study are being used to assess the effect of interference when applying these setups to human motion tracking.