Improving pose estimation accuracy and expanding of visible space of lighting 3D marker in turbid water

Abstract

Aiming at developing underwater battery recharging system, the author developed a docking system using stereo-vision-based visual servoing and a 3D marker. The 3D marker consists of red, green, blue spheres that do not emit the light which is named as a passive marker. Real-time relative pose (position and orientation) estimation was implemented utilizing the 3D model-based matching method and real-time multi-step genetic algorithm (RM-GA). Given the situation that the docking aims for battery recharging in the deep-sea bottom, the pitch-dark and turbid environment should be considered as an inevitable condition for battery recharging. In our previous works, the docking experiments were conducted in the actual sea, having verified the effectiveness of the proposed system using the passive 3D marker in the daytime environment with turbid water condition. Since lighting passive 3D marker by light from the vehicle in turbid water environment results in a situation that the images taken by video cameras set on the vehicle were looked wholly white, some new idea seems to be required. To overcome this difficulty, the newly lighting 3D marker (active 3D maker) has LEDs inside was introduced in the previous work. The main objective of this study is to check the feasibility area of the proposed system for the docking application, comparison of recognition performance using the active and passive 3D marker that was conducted in the simulated pool with the turbid water is focused. And then, the experiment using the active 3D marker in the actual sea has been performed. The experimental results have confirmed that the new active 3D marker with no-lighting from the vehicle could be recognizable in dark and turbid environment than the passive 3D marker with the lighting from the vehicle.