Hesse-Matrix-basierte Qualitätsmanagementsysteme für die Fertigungsindustrie
-
Peng Jieyang
,
Wang Dongkun
Dongkun Wang, a graduate with a Master's degree in Mechanical Engineering from the University of Stuttgart. He is currently pursuing his Ph.D. in the Department of Electronics Engineering at Tsinghua University. His research interests lie in Intelligent Transportation Systems and Data Mining.
Andreas Kimmig is a Master's graduate from Karlsruhe Institute of Technology (KIT), currently serves as a Research Assistant and Ph.D. candidate at KIT. His research focuses on Industrial Big Data and Industry 4.0.
Dr. Armin Roux is an industrial expert at Siemens in Germany, specializing in intelligent fault diagnosis for equipment and Industrial Big Data.
Zusammenfassung
Die Qualitätssicherung ist ein überaus zentrales Thema in der Fertigungsindustrie, da sie unmittelbar mit der Produktqualität und der Kundenzufriedenheit zusammenhängt. Fortschritte in Algorithmen und modernen Kommunikationstechnologien im Kontext von Industrie 4.0 haben dazu beigetragen, dass traditionelle Fertigungsindustrien Deep-Learning-Modelle zur Kontrolle der Produktionsqualität einsetzen. Allerdings stellen industrielle Anwendungen hohe Anforderungen an die Effizienz von Algorithmen. Zudem fehlen in praktischen Anwendungen häufig umfangreiche, gelabelte Daten für das Training von Deep-Learning-Modellen. Um diesen Herausforderungen zu begegnen, haben wir in diesem Artikel ein auf maschinellem Lernen basierendes Modell zur Qualitätserkennung entwickelt. Unser Modell nutzt eine effizientere Hesse-Matrix-Erkennungsmethode, um direkt die lokalen Maxima im Skalenraum des Eingangsbildes zu identifizieren, ohne zahlreiche Gauss-Differenzbilder berechnen zu müssen. Darüber hinaus wenden wir Methoden der Bildverarbeitung an, um die Trainingsdaten zu erweitern, sodass das Modell auch bei geringen Trainingsdatenmengen eine hohe Genauigkeit erreicht. Unsere experimentellen Ergebnisse zeigen, dass das vorgeschlagene Modell die höchste Genauigkeit und Effizienz im Vergleich zu gängigen Methoden aufweist. Abschließend haben wir in diesem Artikel auch eine benutzerfreundliche Schnittstelle für unser Modell erstellt und dieses in das elektronische Kanban der Werkstatt integriert. Unsere empirischen Studien haben ergeben, dass die entwickelten Systeme in der industriellen Praxis anwendbar sind und die Fehlerquote senken sowie die Produktqualität erhöhen können.
Abstract
Quality assurance is a central topic in the manufacturing industry as it directly relates to product quality and customer satisfaction. Advances in algorithms and modern communication technologies in the context of Industry 4.0 have contributed to traditional manufacturing industries utilizing deep learning models for quality control. However, industrial applications impose high demands on the efficiency of algorithms. Furthermore, practical applications often lack extensive, labeled data for the training of deep-learning models. To counter these challenges, we have developed a machine-learning-based quality detection model in this article. Our model uses a more efficient Hesse matrix detection method, to directly identify the local maxima in the scale space of the input image, without having to compute numerous Gauss difference images. Besides, we apply image processing methods to enhance the training data, so that the model achieves high accuracy even with small amounts of training data. Our experimental results show that the proposed model provides the highest accuracy and efficiency compared to other methods. At last, we have also created a user-friendly interface for our model in this article and integrated it into the electronic Kanban of the workshop. Our empirical studies have shown that the developed systems can be applied in industrial practice and can reduce the error rate and improve product quality.
Über die Autoren
Jieyang Peng holds a Ph.D. from Tongji University and is currently a Research Assistant at KIT. His research focuses on Smart Manufacturing and the visualization of Industrial Big Data.
Dongkun Wang, a graduate with a Master's degree in Mechanical Engineering from the University of Stuttgart. He is currently pursuing his Ph.D. in the Department of Electronics Engineering at Tsinghua University. His research interests lie in Intelligent Transportation Systems and Data Mining.
Andreas Kimmig is a Master's graduate from Karlsruhe Institute of Technology (KIT), currently serves as a Research Assistant and Ph.D. candidate at KIT. His research focuses on Industrial Big Data and Industry 4.0.
Dr. Bin Zhang is currently a Senior Engineer at Siemens Research, specializing in the fields of Industry 4.0 and Digital Manufacturing.
Dr. Armin Roux is an industrial expert at Siemens in Germany, specializing in intelligent fault diagnosis for equipment and Industrial Big Data.
Prof. Jivka Ovtcharova serves as the Director of the Institute for Applied Informatics and Formal Description Methods (IMI) at Karlsruhe Institute of Technology (KIT). Holding dual Ph.D. degrees in Computer Science and Mechanical Engineering, her research interests encompass Digital Twins, Industry 4.0, and Digital Manufacturing.
Acknowledgment
Die Forschung wird teilweise durch das National Key R&D Program of China – Construction, Reference Implementation and Verification Platform of Reconfigurable Intelligent Production System (Grant No.2017YFE0101400) unterstützt. Diese Arbeit wurde auch durch das vom EU-Programm H2020 finanzierte Projekt ”DRIMPAC – Unified DR interoperability framework enabling market participation of active energy consumers” unterstützt. (Grant No. 786559).
-
Research ethics: Regarding research ethics, it is hereby stated that this study does not involve any ethical issues.
-
Author contributions: Peng Jieyang: Manuscript preparation, Data analysis and interpretation. Wang Dongkun: Investigation. Andreas Kimmig: Writing - Review & Editing, Validation. Bin Zhang: Data collection. Armin Roux: Conception and design of the study. Jivka Ovtcharova: Supervision, Funding acquisition.
-
Competing interests: Regarding competing interests, the authors declare that they have no financial or non-financial interests that could be perceived as influencing the conduct, analysis, or reporting of this research.
-
Research funding: The research is partially supported by EU H2020 Research and Innovation Program under the Marie Sklodowska-Curie Grant Agreement (Project-DEEP, Grant No. 101109045). The research is also partially funded by the German Federal Ministry of Education and Research (BMBF) (project AITT - AI-assisted Technology Transfer, No. 03LB3058B).
-
Data availability: The raw data can be obtained on request from the corresponding author.
References
[1] S. Archambeau, “What is Your Company’s Cost of Poor Quality? Tools for Calculating and Reducing It,” Quality Digest, 2004.Search in Google Scholar
[2] A. Srinivasan and B. Kurey, “Creating a culture of quality,” Harv. Bus. Rev., vol. 92, no. 4, pp. 23–25, 2014.Search in Google Scholar
[3] T. Pfeifer, W. Reissiger, and C. Canales, “Integrating six sigma with quality management systems,” TQM Mag., vol. 16, no. 4, pp. 241–249, 2004. https://doi.org/10.1108/09544780410541891.Search in Google Scholar
[4] C. Y. Lee and X. Zhou, “Quality management and manufacturing strategies in China,” Int. J. Qual. Reliab. Manag., vol. 17, no. 8, pp. 876–899, 2000. https://doi.org/10.1108/02656710010325101.Search in Google Scholar
[5] A. Anvari, et al.., “A study on total quality management and lean manufacturing: through lean thinking approach,” World Appl. Sci. J., vol. 12, no. 9, pp. 1585–1596, 2011.Search in Google Scholar
[6] F. J. Lloréns-Montes, V. J. García-Morales, and A. J. Verdú-Jover, “Flexibility and quality management in manufacturing: an alternative approach,” Prod. Plan. Control, vol. 15, no. 5, pp. 525–533, 2004. https://doi.org/10.1080/09537280410001726339.Search in Google Scholar
[7] A. Zonnenshain and R. S. Kenett, “Quality 4.0—the challenging future of quality engineering,” Qual. Eng., vol. 32, no. 4, pp. 614–626, 2020. https://doi.org/10.1080/08982112.2019.1706744.Search in Google Scholar
[8] H. Jung, J. Jeon, D. Choi, and J.-Y. Park, “Application of machine learning techniques in injection molding quality prediction: implications on sustainable manufacturing industry,” Sustainability, vol. 13, no. 8, p. 4120, 2021. https://doi.org/10.3390/su13084120.Search in Google Scholar
[9] M. Javaid, A. Haleem, R. P. Singh, S. Rab, and R. Suman, “Exploring impact and features of machine vision for progressive industry 4.0 culture,” Sens. Int., vol. 3, 2022, Art. no. 100132. https://doi.org/10.1016/j.sintl.2021.100132.Search in Google Scholar
[10] R. Miao, et al.., “Online inspection of narrow overlap weld quality using two-stage convolution neural network image recognition,” Mach. Vis. Appl., vol. 32, pp. 1–14, 2021. https://doi.org/10.1007/s00138-020-01158-2.Search in Google Scholar
[11] Y.-J. Cha, W. Choi, and O. Büyüköztürk, “Deep learning-based crack damage detection using convolutional neural networks,” Comput.-Aided Civ. Infrastruct. Eng., vol. 32, no. 5, pp. 361–378, 2017. https://doi.org/10.1111/mice.12263.Search in Google Scholar
[12] Y. Fu, et al.., “Machining vibration states monitoring based on image representation using convolutional neural networks,” Eng. Appl. Artif. Intell., vol. 65, pp. 240–251, 2017. https://doi.org/10.1016/j.engappai.2017.07.024.Search in Google Scholar
[13] Y.-J. Cha, W. Choi, G. Suh, S. Mahmoudkhani, and O. Büyüköztürk, “Autonomous structural visual inspection using region-based deep learning for detecting multiple damage types, Computer-Aided Civil and Infrastructure,” Engineering, vol. 33, no. 9, pp. 731–747, 2018. https://doi.org/10.1111/mice.12334.Search in Google Scholar
[14] D. He, K. Xu, and P. Zhou, “Defect detection of hot rolled steels with a new object detection framework called classification priority network,” Comput. Ind. Eng., vol. 128, pp. 290–297, 2019. https://doi.org/10.1016/j.cie.2018.12.043.Search in Google Scholar
[15] S. Klus and P. Gelß, “Tensor-based algorithms for image classification,” Algorithms, vol. 12, no. 11, p. 240, 2019. https://doi.org/10.3390/a12110240.Search in Google Scholar
[16] D. G. Lowe, “Distinctive image features from scale-invariant keypoints,” Int. J. Comput. Vis., vol. 60, pp. 91–110, 2004. https://doi.org/10.1023/b:visi.0000029664.99615.94.10.1023/B:VISI.0000029664.99615.94Search in Google Scholar
[17] M. D. Prasetio, M. H. Rifai, and R. Y. Xavierullah, “Design of defect classification on clay tiles using support vector machine (svm),” in 2020 6th International Conference on Interactive Digital Media (ICIDM), IEEE, 2020, pp. 1–6.10.1109/ICIDM51048.2020.9339642Search in Google Scholar
[18] O. Augereau, N. Journet, A. Vialard, and J.-P. Domenger, “Improving classification of an industrial document image database by combining visual and textual features,” in 2014 11th IAPR International Workshop on Document Analysis Systems, IEEE, 2014, pp. 314–318.10.1109/DAS.2014.44Search in Google Scholar
[19] L. D. Medus, M. Saban, J. V. Frances-Villora, M. Bataller-Mompean, and A. Rosado-Muñoz, “Hyperspectral image classification using cnn: application to industrial food packaging,” Food Control, vol. 125, 2021, Art. no. 107962. https://doi.org/10.1016/j.foodcont.2021.107962.Search in Google Scholar
[20] A. Azulay and Y. Weiss, “Why do deep convolutional networks generalize so poorly to small image transformations?” arXiv preprint arXiv:1805.12177, 2019.Search in Google Scholar
© 2024 Walter de Gruyter GmbH, Berlin/Boston
Articles in the same Issue
- Frontmatter
- Editorial
- Vorwort 2025
- Survey
- Information Circularity Assistance based on extreme data
- Methods
- Subjective task-load influences anthropomorphism during cooperative human and robot hand movements
- Applications
- Concept study of an autonomous aerial mobile network relay for pre-hospital emergency care
- Challenges and requirements of AI-based waste water treatment systems
- Modeling propagation competition between hostile influential groups using opinion dynamics
- Hesse-Matrix-basierte Qualitätsmanagementsysteme für die Fertigungsindustrie
Articles in the same Issue
- Frontmatter
- Editorial
- Vorwort 2025
- Survey
- Information Circularity Assistance based on extreme data
- Methods
- Subjective task-load influences anthropomorphism during cooperative human and robot hand movements
- Applications
- Concept study of an autonomous aerial mobile network relay for pre-hospital emergency care
- Challenges and requirements of AI-based waste water treatment systems
- Modeling propagation competition between hostile influential groups using opinion dynamics
- Hesse-Matrix-basierte Qualitätsmanagementsysteme für die Fertigungsindustrie
