Shape-Based Object Pose Estimation for Collaborative Robot with 2-Finger Gripper

A new shape-based object pose estimation for collaborative robot with two-finger gripper has been proposed here. This method uses the smallest enclosing shapes target and takes advantage of symmetrical characteristic to determine gripping direction gripper. First, interesting pick place will be detected segmented. Second, segmented is transformed into its convex form. Next, are approximated best-fit shape selected. Last, calculated according symmetry.