Intelligent object detection and robot manipulation

Abstract

As the need for developing advanced manufacturing based on 3D object detection, new techniques are developed and optimized to fulfill the requirements. This project involves 3D object detection with input of 2D, 3D images and point cloud. The pose estimation of the object for ease of capture and handling will be studied. Robot arms control method was applied for the corresponding robot manipulation on targetobject for capturing and transferring to new position. The 3D object detection system which was called SKL-IOTSC 3D vision guidance system was invented by State Key Laboratory for the Internet of Things for Smart City (University of Macau) (Figure 1) a few years ago. SKL-IOTSC 3D vision guidance system involved structured light, point cloud and surface reconstruction to detected target object and transferring to desire position by UR5 robot arm. This project methodology and result were based on SKL-IOTSC 3D vision guidance system. The purpose of this report is to detected target object, analyzed point cloud artifacts and measured accuracy of SKL-IOTSC 3D vision guidance system.