Pick and place industrial robot controller with computer vision

Abstract

In these days, most of the robots operate in a precoded cycle, to the best, the cycle waits for a few input sensors to further along cycle. When we compare this with the way a human being does the same job, we have a lot more flexibility as the path to be taken by our hand is formulated for each cycle depending on the environment under robot. This paper gives an idea about finding the x, y coordinate of object. This is found by converting pixel co-ordinate into real world co-ordinate with the help of 2D transformation. We incorporate robotic arm controller based on location and orientation of object. Fundamentally the problem taken in this work is to pick and place objects which are placed under vision sensor. Computer vision is the visual discernment segment of an eager motivation to copy human insight and to enrich robots with savvy conduct. This work has been carried under the guidance of Castalia Research lab, Pune.