부산대학교 도서관

To extend the persistence time of an underwater operation of AUVs in the sea, many studies have been performed worldwide. The docking function takes place as an important role not only for battery recharging but also for other advanced applications. Therefore, we developed the visual-servoing-based underwater vehicle using 3D perception based move on sensing (3D-MoS) system with dual-eye cameras and 3D marker for the docking operation. The vision-based underwater vehicle cannot avoid the turbidity and low illumination in the deep sea. In the present study, the active/lighting marker was newly designed and constructed to improve the recognizing ability of the proposed system for the real-time 3D pose estimation. The experiments were conducted in a simulated pool against different turbidity levels and illumination variations by adjusting the LED's brightness of the active marker in both day and night environments. This paper presents the analyses on the relations of the currents of LED installed inside each ball of the 3D marker and recognition results under turbidity and changing illumination conditions. According to the experimental results, the optimum current was chosen for the docking.