A Calibration-Free Visual System for Device Pose Estimation

This paper focuses on device pose estimation in automation processes and an innovative calibration-free visual system is proposed. We introduce a novel solution employing three orthogonal cameras to form an efficient relocalization system. Through the analysis of scene images captured by these cameras, we determine the position of cameras and their orientation in three-dimensional space. By processing the scene images captured by cameras, we extract crucial position and pose information of cameras in the spatial domain. Our method boasts notable advantages in terms of efficiency and speed.