Document Server@UHasselt >
Education >
Faculty of Engineering Technology >
Master theses >

Please use this identifier to cite or link to this item: http://hdl.handle.net/1942/24616

Title: Detection of Lint by Using Machine Vision
Authors: Filtjens, Benjamin
Advisors: LEEN, Geert
LEHTO, Raine
Issue Date: 2017
Publisher: UHasselt
Abstract: This thesis, commissioned by Häme University of Applied Sciences, researches the possibility of detecting lint by using machine vision. Due to the small particle size and high movement speed of the lint, various issues occur. Firstly, to detect the small lint particles a sufficient resolution is required. Secondly, since the lint has a high movement speed a high framerate is required to fully represent all the lint passing by. Lastly, a short exposure time is required to prevent inaccuracy due to motion blur. The goals of this thesis are to research the most optimal machine vision components, if the hardware currently available can detect the small particles with a sufficient framerate and a method to prevent motion blur. The most optimal components were found by performing a literature study. Calculations were made to test if the currently available hardware can fulfil the goals. A colleague created a short duration strobe light to prevent motion blur. Lastly, a practical test setup and MATLAB program were created to verify the theoretical conclusions and detect the lint. The strobe light uses four high power white LEDs with a flash duration of one microsecond. The calculations have concluded that the currently available hardware is capable of fully representing the lint passing by at a minimum particle size of 45 microns. Analyses of the MATLAB program verified that the theoretical calculations were correct.
Notes: master in de industriële wetenschappen: energie-automatisering
URI: http://hdl.handle.net/1942/24616
Category: T2
Type: Theses and Dissertations
Appears in Collections: Master theses

Files in This Item:

Description SizeFormat
N/A2.54 MBAdobe PDF
N/A637.59 kBAdobe PDF

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.