主催: The Japan Society of Mechanical Engineers
会議名: ロボティクス・メカトロニクス 講演会2018
開催日: 2018/06/02 - 2018/06/05
Visual servoing is one of the methods to control robots. By incorporating visual information obtained from the installed visual sensor into the feedback loop, it is expected to enable robots to work in changing environments or unknown environment. For a visual servoing system, it is difficult to quickly switch the tracking object for different tasks. Using a dual hand-eye cameras manipulator, authors have proposed photo-model-based recognition method for handling different objects with static pictures. In this paper, the recognition method is used for the pose tracking of different objects with dynamic pictures. The visual servoing experimental results demonstrate the proposed recognition method are feasible, flexible and effective.