| Summary: | 碩士 === 國立臺灣科技大學 === 機械工程系 === 101 === This dissertation studies and implements a control scheme of robot manipulator to perform a grasping task using visual servoing techniques. Visual servoing is a multidiscipline research topic with several subjects being discussed in this thesis including camera model and calibration, image correspondence, pose estimation and the well-known visual servoing control scheme image-based visual servoing (IBVS). At initial control phase and first coarse approach, tabletop segmentation from Robot Operating System (ROS) inspires the idea of a simpler adaptation, using background segmentation, applied for 2D processing. Later, at IBVS control phase with fine approach, a simple object detection capability is applied to recognize suitable grasping poses as desired image features for IBVS control.
|
|---|
