Grasping novel objects with depth segmentation

We consider the task of grasping novel objects and cleaning fairly cluttered tables with many novel objects. Recent successful approaches employ machine learning algorithms to identify points on the scene that the robot should grasp. In this paper, we show that the task can be significantly simplified by using segmentation, especially with depth information. A supervised localization method is employed to select graspable segments. We also propose a shape completion and grasp planner method which takes partial 3D information and plans the most stable grasping strategy. Extensive experiments on our robot demonstrate the effectiveness of our approach.

Authors

• Deepak Rao
• Quoc V. Le
• Thanathorn Phoka
• Morgan Quigley
• Attawith Sudsang
• Andrew Y. Ng

Keywords

• Grasping
• Three dimensional displays
• Image segmentation
• Robot sensing systems
• Shape
• Pixel
• Depth Segment
• Recent Approaches
• Depth Information
• 3D Information
• Support Vector Machine
• Local Features
• Normal Vector
• Pixel Intensity
• Point Cloud
• Bounding Box
• Geometric Features
• Segmentation Algorithm
• 3D Data
• Depth Data
• Threshold Function
• 3D Point Cloud
• Objects In The Scene
• Triangular Mesh
• Visible Images
• Robotic Hand
• 3D Bounding Box
• Visible Light Images
• Complete 3D
• Local Mesh
• Degrees Of Freedom
• Radial Basis Function
• Robot Manipulator
• Collision Detection
• Active Sensors
• 3D Mesh