This paper presents an automated machine vision system for the quality inspection of blanked metal components by the analysis of product images. The binary classification to identify defective samples among the compliant ones is carried out by means of two different supervised machine learning algorithms, a Sequential Neural Network (SNN) and a Support Vector Machine (SVM). The features used by the algorithms are extracted from the images using a novel approach, that exploits some geometrical characteristics of the product. Indeed, the spatial features of the target object in the images are processed as they represent a time series signal and hence features are extracted and exploited as input data for the classification task.

Automated Vision Inspection of Critical Steel Components based on Signal Analysis Extracted form Images

Brambilla P.;Chiariotti P.;Tarabini M.
2022-01-01

Abstract

This paper presents an automated machine vision system for the quality inspection of blanked metal components by the analysis of product images. The binary classification to identify defective samples among the compliant ones is carried out by means of two different supervised machine learning algorithms, a Sequential Neural Network (SNN) and a Support Vector Machine (SVM). The features used by the algorithms are extracted from the images using a novel approach, that exploits some geometrical characteristics of the product. Indeed, the spatial features of the target object in the images are processed as they represent a time series signal and hence features are extracted and exploited as input data for the classification task.
2022
Proceedings of the 2022 IEEE International Workshop on Metrology for Industry 4.0 and IoT, MetroInd 4.0 and IoT 2022
978-1-6654-1093-9
File in questo prodotto:
File Dimensione Formato  
Automated_Vision_Inspection_of_Critical_Steel_Components_based_on_Signal_Analysis_Extracted_form_Images.pdf

Accesso riservato

: Publisher’s version
Dimensione 1.95 MB
Formato Adobe PDF
1.95 MB Adobe PDF   Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11311/1222071
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 4
  • ???jsp.display-item.citation.isi??? 1
social impact