Model-based robotic visual servoing

This paper presents a model-based approach for visual tracking and servoing in robotics. We propose active deformable models as an effective way for tracking a rigid or semi-rigid (possibly partially occluded) object in movement within the manipulator's workspace. Deformable models simulate the behavior of elastic edges or meshes in real-time. Such models can be designed to adhere to the contour of a rigid or semi-rigid object as defined by the image gradient. Because these models retain a 'memory' of the shape and orientation of the object they track, they can be used effectively in model-based robotic visual servoing. Preliminary experimental results show that the deformable models help us achieve robust tracking even when the target is partially occluded.