Visual control of a robotic arm

Abstract

Integration of robot control with computer vision is an emerging technology. The ability to capture a real world scenario, filter out a specific object, and then localise it with respect to a specific point opens up many avenues for future technologies. In this paper, computer vision is integrated with a 6-DOF robotic arm. The aim is to develop a methodology for controlling a robotic arm through visual interaction. The investigations carried out include mathematical and geometrical representation of the physical manipulator, object detection utilising a colour segmentation approach, establishing communications from the command line of the visual recognition program with an Arduino and development of a user orientated software package that integrates the two key objectives.