An Approach for Object Picking using Correlation analysis and Robotic Manipulation

Abstract

The aim of the present work is to pick and place an object with the help of robotic manipulator using computer vision. Real time object detection is achieved using a camera to capture the image in the region of interest (ROI). 2-D correlation between the object and captured image, along with image processing techniques, locates the object position. The pixel coordinates are then converted to real time distance coordinates. The object location is conveyed to the robotic arm with the help of software instructions and required operation of pick/place is performed. The object is successfully located and grabbed by the robotic arm, provided the relative position of the robotic arm and the camera is fixed. The method uses two-dimensional image analysis in place of three-dimensional object detecting techniques. Hence the proposed algorithm is simple, requires less memory space, does not require any filter, reduces the instruction set and has high accuracy.