Author(s)

Haikuan Du

Corresponding Author

Haikuan Du

ABSTRACT

Object grabbing by robotic arms is a research hotspot in current human-computer interaction research. This paper first introduces the overall process framework of the robotic arm grasping system, and then begins to introduce the specific operation process or specific implementation method of each part of the system. The method in this paper first uses a depth camera to reconstruct 3D objects. Then, before using 3D point cloud matching to solve the object pose, the current popular and highly accurate 2D object detection method is used to filter scene point clouds. In this way, it makes the speed of subsequent point cloud matching greatly improved, which also improves the speed and accuracy of the entire grasping system.

KEYWORDS

3D Object Matching; Object Detection; Robotic Arm Grasping System