Stereo and Mono Vision Sorting of Known Objects and Grasping Using Industrial Robot Arm

Abstract

This thesis presents the development and experimental implementation of an automated system to check the dimensions and grasp a classified object using an industrial robot arm. The system consists of a 6 DOF Mitsubishi robot arm, PC, customized vision system and also Ethernet communication to transfer data from the computer to robot controller to enable robot arm to grip the target product. The vision system begins with a stream video and sends the video to vision system algorithm that analyzes the video frame by frame. The vision system algorithm extracts the data from the video and obtains the three-dimensional coordinate of a point on the surface of the classified object and sends this location to the robot controller via Ethernet communication, and also checks the dimensions of the object. The robot controller receives this location from the PC then converts the robot pose into joint angles using inverse kinematics algorithm. The vision system is based on a single camera or a stereo camera. The task of the vision system is to obtain 3D coordinates of a point on the surface of the object, classifies between different objects based on a shape detection algorithm and also checks the dimensions of the object