Analysis Of Shape Recognition Using Robotic Fingers By Image Processing Technique For Object Manipulation

Musa, Muhamad Aqmal (2019) Analysis Of Shape Recognition Using Robotic Fingers By Image Processing Technique For Object Manipulation. Project Report. UTeM, Melaka, Malaysia. (Submitted)

Text (24 Pages) Analysis Of Shape Recognition Using Robotic Fingers By Image Processing Technique For Object Manipulation.pdf - Submitted Version Download (376kB)

Abstract

In line with the advancement of technologies, present tactile sensing based gripping technologies have shown improvements in interactive force gripping and identification of object. However, the precision of shape recognition using tactile sensors on gripper is unsatisfied. The main objective for this study is to design and develop a system for identifying the contact points of different object shapes and analyze the angular motion of fingers movement as well as accuracy of object grasped. The design of gripper involves two robotic fingers with two degree of freedom. It implements basic image processing techniques such as gray-scale, threshold and edge detection as the approach to determine the vertices on the object image for recognizing the object’s shape. Plus, the contact points on the object surface that will be touched by the gripper’s fingers define the orientations of the joints by using inverse kinematic equation. The effectiveness of this vision system required large size of targeted object compared to smaller size. These analyses obtained 2.05 and 2.0 pixels per millimeter which have been converted from pixel coordination to the real coordination for circle object and square object respectively. The accuracy on the orientations calculated in actual value and pixel value were just in slightly different which were 97.54% (Link 1) and 97.4% (Link 2) for grasping circle object while 97.48% (Link 1) and 93.8% (Link 2) for grasping the square object.

Actions (login required)

View Item