Deep learning-supported machine vision-based hybrid system combining inhomogeneous 2D and 3D data for the identification of surface defects

Machine vision systems for automatic defect detection commonly adopt 2D image-based systems or 3D laser triangulation systems. 2D and 3D systems present opposite advantages and disadvantages depending on the typology and position of defects to be detected. When the variety of defects is large, none...

Full description

Saved in:
Bibliographic Details
Main Authors: Giorgio Cavaliere, Oswald Lanz, Yuri Borgianni, Enrico Savio
Format: Article
Language:English
Published: Taylor & Francis Group 2024-12-01
Series:Production and Manufacturing Research: An Open Access Journal
Subjects:
Online Access:https://www.tandfonline.com/doi/10.1080/21693277.2024.2378199
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846131235272785920
author Giorgio Cavaliere
Oswald Lanz
Yuri Borgianni
Enrico Savio
author_facet Giorgio Cavaliere
Oswald Lanz
Yuri Borgianni
Enrico Savio
author_sort Giorgio Cavaliere
collection DOAJ
description Machine vision systems for automatic defect detection commonly adopt 2D image-based systems or 3D laser triangulation systems. 2D and 3D systems present opposite advantages and disadvantages depending on the typology and position of defects to be detected. When the variety of defects is large, none of them performs defect detection accurately. To overcome this limitation, this paper illustrates a hybrid Deep Learning-supported system where the 2D- and 3D-generated data are juxtaposed and analyzed contextually. Anomaly scores are subsequently determined to distinguish suitable and uncompliant parts. The implementation of the hybrid system allowed the identification of defective parts in an aluminium die-cast component with an accuracy concerning true positives of over 95% by comparing the system outputs with human defect detection. The inspection time was reduced by approximately 20% if compared, once again, with the same activities performed by humans.
format Article
id doaj-art-958dfa97134b4c4d832ea01b3fa3be49
institution Kabale University
issn 2169-3277
language English
publishDate 2024-12-01
publisher Taylor & Francis Group
record_format Article
series Production and Manufacturing Research: An Open Access Journal
spelling doaj-art-958dfa97134b4c4d832ea01b3fa3be492024-12-09T15:50:02ZengTaylor & Francis GroupProduction and Manufacturing Research: An Open Access Journal2169-32772024-12-0112110.1080/21693277.2024.2378199Deep learning-supported machine vision-based hybrid system combining inhomogeneous 2D and 3D data for the identification of surface defectsGiorgio Cavaliere0Oswald Lanz1Yuri Borgianni2Enrico Savio3Faculty of Engineering, Free University of Bozen-Bolzano, Piazza Università, Bolzano (BZ), ItalyFaculty of Engineering, Free University of Bozen-Bolzano, Piazza Università, Bolzano (BZ), ItalyFaculty of Engineering, Free University of Bozen-Bolzano, Piazza Università, Bolzano (BZ), ItalyDepartment of Industrial Engineering, University of Padua, Padua (PD), ItalyMachine vision systems for automatic defect detection commonly adopt 2D image-based systems or 3D laser triangulation systems. 2D and 3D systems present opposite advantages and disadvantages depending on the typology and position of defects to be detected. When the variety of defects is large, none of them performs defect detection accurately. To overcome this limitation, this paper illustrates a hybrid Deep Learning-supported system where the 2D- and 3D-generated data are juxtaposed and analyzed contextually. Anomaly scores are subsequently determined to distinguish suitable and uncompliant parts. The implementation of the hybrid system allowed the identification of defective parts in an aluminium die-cast component with an accuracy concerning true positives of over 95% by comparing the system outputs with human defect detection. The inspection time was reduced by approximately 20% if compared, once again, with the same activities performed by humans.https://www.tandfonline.com/doi/10.1080/21693277.2024.2378199Surface defects analysishybrid 2D and 3D systemmachine vision systemsdeep learninghigh-pressure die-casting
spellingShingle Giorgio Cavaliere
Oswald Lanz
Yuri Borgianni
Enrico Savio
Deep learning-supported machine vision-based hybrid system combining inhomogeneous 2D and 3D data for the identification of surface defects
Production and Manufacturing Research: An Open Access Journal
Surface defects analysis
hybrid 2D and 3D system
machine vision systems
deep learning
high-pressure die-casting
title Deep learning-supported machine vision-based hybrid system combining inhomogeneous 2D and 3D data for the identification of surface defects
title_full Deep learning-supported machine vision-based hybrid system combining inhomogeneous 2D and 3D data for the identification of surface defects
title_fullStr Deep learning-supported machine vision-based hybrid system combining inhomogeneous 2D and 3D data for the identification of surface defects
title_full_unstemmed Deep learning-supported machine vision-based hybrid system combining inhomogeneous 2D and 3D data for the identification of surface defects
title_short Deep learning-supported machine vision-based hybrid system combining inhomogeneous 2D and 3D data for the identification of surface defects
title_sort deep learning supported machine vision based hybrid system combining inhomogeneous 2d and 3d data for the identification of surface defects
topic Surface defects analysis
hybrid 2D and 3D system
machine vision systems
deep learning
high-pressure die-casting
url https://www.tandfonline.com/doi/10.1080/21693277.2024.2378199
work_keys_str_mv AT giorgiocavaliere deeplearningsupportedmachinevisionbasedhybridsystemcombininginhomogeneous2dand3ddatafortheidentificationofsurfacedefects
AT oswaldlanz deeplearningsupportedmachinevisionbasedhybridsystemcombininginhomogeneous2dand3ddatafortheidentificationofsurfacedefects
AT yuriborgianni deeplearningsupportedmachinevisionbasedhybridsystemcombininginhomogeneous2dand3ddatafortheidentificationofsurfacedefects
AT enricosavio deeplearningsupportedmachinevisionbasedhybridsystemcombininginhomogeneous2dand3ddatafortheidentificationofsurfacedefects