TU Darmstadt / ULB / TUbiblio

Automated quality control of vacuum insulated glazing by convolutional neural network image classification

Riedel, Henrik ; Mokdad, Sleheddine ; Schulz, Isabell ; Kocer, Cenk ; Rosendahl, Philipp Laurens ; Schneider, Jens ; Kraus, Michael A. ; Drass, Michael (2022)
Automated quality control of vacuum insulated glazing by convolutional neural network image classification.
In: Automation in Construction, 135
doi: 10.1016/j.autcon.2022.104144
Article, Bibliographie

Abstract

Vacuum insulated glazing (VIG) is a highly thermally insulating window technology, which boasts an extremely thin profile and lower weight as compared to gas-filled insulated glazing units of equivalent performance. The VIG is a double-pane configuration with a submillimeter vacuum gap maintained by small pillars positioned in between the panes, which can damage the glass during manufacturing, transportation and installation. For the purpose of automatically classifying the damage, we have developed, trained, and tested a deep learning model using convolutional neural networks. We employ the state-of-the-art methods Grad-CAM and Score-CAM of explainable Artificial Intelligence (XAI) to provide an understanding of the internal mechanisms and were able to show that our classifier outperforms ResNet50V2 for identification of crack locations and geometry. Further analysis of our model's predictive capabilities demonstrates its superiority over state-of-the-art ResNet models in terms of convergence speed, accuracy, precision at 100 recall and AUC for ROC.

Item Type: Article
Erschienen: 2022
Creators: Riedel, Henrik ; Mokdad, Sleheddine ; Schulz, Isabell ; Kocer, Cenk ; Rosendahl, Philipp Laurens ; Schneider, Jens ; Kraus, Michael A. ; Drass, Michael
Type of entry: Bibliographie
Title: Automated quality control of vacuum insulated glazing by convolutional neural network image classification
Language: English
Date: 22 January 2022
Publisher: Elsevier
Journal or Publication Title: Automation in Construction
Volume of the journal: 135
DOI: 10.1016/j.autcon.2022.104144
URL / URN: https://www.sciencedirect.com/science/article/pii/S092658052...
Abstract:

Vacuum insulated glazing (VIG) is a highly thermally insulating window technology, which boasts an extremely thin profile and lower weight as compared to gas-filled insulated glazing units of equivalent performance. The VIG is a double-pane configuration with a submillimeter vacuum gap maintained by small pillars positioned in between the panes, which can damage the glass during manufacturing, transportation and installation. For the purpose of automatically classifying the damage, we have developed, trained, and tested a deep learning model using convolutional neural networks. We employ the state-of-the-art methods Grad-CAM and Score-CAM of explainable Artificial Intelligence (XAI) to provide an understanding of the internal mechanisms and were able to show that our classifier outperforms ResNet50V2 for identification of crack locations and geometry. Further analysis of our model's predictive capabilities demonstrates its superiority over state-of-the-art ResNet models in terms of convergence speed, accuracy, precision at 100 recall and AUC for ROC.

Uncontrolled Keywords: Deep learning, Classification, Vacuum insulating glass, Object detection, Data augmentation, Explainable AI, Grad-CAM, Score-CAM
Additional Information:

Artikel ID: 104144

Divisions: 13 Department of Civil and Environmental Engineering Sciences
13 Department of Civil and Environmental Engineering Sciences > Institute für Structural Mechanics and Design
13 Department of Civil and Environmental Engineering Sciences > Institute für Structural Mechanics and Design > Structural Engineering
Date Deposited: 25 Jan 2022 07:08
Last Modified: 07 Feb 2022 10:26
PPN:
Export:
Suche nach Titel in: TUfind oder in Google
Send an inquiry Send an inquiry

Options (only for editors)
Show editorial Details Show editorial Details