Vizijski sustav za automatsku kontrolu kvalitete uložaka grebenaste sklopke

master's thesis

• Title (english): Machine vision system for automated quality control of rotary switch inserts
• Author: Emanuela Rešetar
• Mentor: Tomislav Staroveški (mentor)
• Committee member: Josip Kasać (predsjednik povjerenstva)
• Committee member: Zoran Kunica (član povjerenstva)
• Committee member: Ivica Garašić (član povjerenstva)
• Committee member: Tomislav Staroveški (član povjerenstva)
• Granter: University of Zagreb; Faculty of Mechanical Engineering and Naval Architecture; Zagreb
• Defense date and country: 2020-03-12, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Mechanical Engineering

Abstract

Rad se bavi razvojem vizijskog sustava za prepoznavanje i lokalizaciju uložaka grebenaste sklopke. Predstavljene su dvije algoritamske metode prikladne za robusno prepoznavanje. Prva algoritamska metoda koristi Houghovu kružnu transformaciju na binarnim slikama kako bi se locirali uzorci, koji se zatim kategoriziraju analizom intenziteta obodnih piksela. Druga algoritamska metoda temelji se na segmentaciji slike i analizi bloba pri čemu se koriste momenti za lokalizaciju i analiza konture za prepoznavanje. Oba pristupa testirana su na velikom broju slika različite složenosti (u rasponu od jednostavnih slika bez preklapanja uzoraka do slika sa znatnim preklapanjem uzoraka). Rezultati pokazuju kako su obje algoritamske metode prikladne za robusno prepoznavanje i lokalizaciju uzoraka. Međutim, prvi algoritam pokazao se bolji u pogledu vremena odziva i računske intenzivnosti.

Abstract (english)

This thesis presents the development of a machine vision system for the classification and localization of rotary switch inserts. In order to achieve robust results, two approaches were investigated. In the first approach, Hough transform for circles was used on binarized edge images in order to localize inserts, which were then classified by analyzing pixel intensities across the insert perimeter. The second approach is based on the image segmentation and blob analysis in which moments were used for localization and contour analysis for classification. Both approaches were tested on a large number of images with different scene complexities (ranging from simple images with no insert overlapping to images with considerable insert overlapping). Results show that both approaches can be used for this application, however, the first approach had a lower response time and was less computationally intensive.

• Keywords: strojni vid; računalni vid; procesiranje slike; Houghova transformacija; detekcija kontura; OpenCV
• Keywords (english): Machine vision; Computer Vision; Image processing; Hough Transformation; Detection of Contours; OpenCV
• Language: croatian
• URN:NBN: urn:nbn:hr:235:068446
• Study programme: Title: Mechanical Engineering; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka strojarstva (magistar/magistra inženjer/inženjerka strojarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2020-12-02 10:52:57