A three dimensional (3D) vision based defect inspection system for gluing application

Harun, Mohamad Haniff (2013) A three dimensional (3D) vision based defect inspection system for gluing application. Masters thesis, Universiti Teknikal Malaysia Melaka.

[img]
Preview
Text (24 Pages)
Complete_Thesis_(Mohamad_Haniff_bin_Harun)_-_with_Sign.pdf - Submitted Version

Download (7MB)
[img] Text (Full Text)
A three dimensional (3D) vision based defect inspection system for gluing application.pdf - Submitted Version
Restricted to Registered users only

Download (7MB)

Abstract

A Robot Vision System (RVS) is an adaptive and dynamic system that caters to a wide range of jobs where each involves a set of operations required to be done at a predetermined workstation. This research is focused on the development of a vision system to be integrated with KUKA arm robot. Pyramid object is used as a complimentary of the windscreen car as a model. It developed using plain cardboard with dimension of 15cm x 15cm. 2D matching application introduced to identify the characteristic of the object used in the system using CCD camera. Object used must be trained in training phase to create object template and used again in recognition phase for object classification. Then, two CCD cameras are used; placed at the top and front of the object to extract object’s edge location using Harris Point. Data extracted from it are used to find 3D coordination of each edge. Equation of straight line mostly used in this method to identify x, y and z coordinates. Data obtained from the system then used to give instruction to KUKA arm robot for gluing purposes. Pixel coordinates must be converted to robot coordinates for easier understanding by the robot. Three types of defect are trained as model templates and save to the memory known as bumper, gap and bubble defect. Each defect has special characteristic. Inspection system developed to identify problems occurs in gluing process. Template matching method used to call model trained in training phase to identify the uncertainties. Each defect occurs comes with its coordinate’s information for correction. Correction of defect consists of two phase; 1st CoD where correction is completed in first time and 2nd CoD where correction still need to be completed after the first correction. Data for all the process are recorded to prove that this algorithm made improvement with the previous research.

Item Type: Thesis (Masters)
Uncontrolled Keywords: Robotics, Computer vision, Robot vision
Subjects: T Technology > T Technology (General)
T Technology > TJ Mechanical engineering and machinery
Divisions: Library > Tesis > FKE
Depositing User: Mohamad Haniff Harun
Date Deposited: 12 Mar 2014 06:30
Last Modified: 20 Apr 2022 10:21
URI: http://eprints.utem.edu.my/id/eprint/11219
Statistic Details: View Download Statistic

Actions (login required)

View Item View Item