Nur Hazirah, Ismail (2015) Shape Based Matching Gluing Defect Using Gaussian Filtering. Project Report. UTeM, Melaka, Malaysia. (Submitted)
![]() |
Text (24 pages)
Shape Based Matching Gluing Defect Using Gaussian Filtering 24 Pages.pdf - Submitted Version Download (6MB) |
Abstract
This research project is regarding the application of vision algorithm which is related with the images processing process in order to identify defectives area of a product. This paper proposes an efficient defect detection and classification technique to detect a class of defects in gluing application such as gap, bubble, bumper, and edge defect that can have ensured the better quality of product in manufacturing process as well as production rate in industrial automation for gluing process. In determining better image processing, Gaussian filtering method is proposed because this filter is widely used in image processing as it can be used to smooth textures for image segmentation, removing noise, and keep more texture details. The main focus of this paper is on the shape based matching properties method where it mainly focus on identifying defects that occur on the product by finding the object based on the single image and locate them with sub pixel accuracy. The algorithm consists of two phases which are training phase and the recognition phase. In this project, creating of region of interest in important region of object is discussed and two-dimensional (2D) shape matching was chosen to be used to provide information of width ex-coordinate) and length (y-coordinate).
Item Type: | Final Year Project (Project Report) |
---|---|
Uncontrolled Keywords: | Image processing, Computer vision |
Subjects: | T Technology > TA Engineering (General). Civil engineering (General) |
Divisions: | Library > Final Year Project > FTK |
Depositing User: | Ahmad Tarmizi Abdul Hadi |
Date Deposited: | 01 Apr 2016 07:39 |
Last Modified: | 01 Apr 2016 07:39 |
URI: | http://digitalcollection.utem.edu.my/id/eprint/16048 |
Actions (login required)
![]() |
View Item |