Improving Quality Control in Casting Manufacturing Using Vision Transformers
Abstract
Quality control is vital in modern manufacturing to ensure product reliability and competitiveness. This paper addresses defect detection in casting discs for submersible pump impellers using automated visual inspection. We propose a method based on Vision Transformers (ViT), which leverage selfattention to learn visual patterns effectively. Real and simulated experiments showed high accuracy (99.22%) and strong robustness to image noise, maintaining 95.45% and 98.28% accuracy under Gaussian and Salt-and-Pepper noise. The results confirm the method’s reliability and potential to optimize the casting process.References
Dabhi, R. (2020). Casting product image data for quality inspection. [link]. Accessed: 2024-05-08.
De Souza Gil, E., De Abreu Dias, L. M., De Souza Brito, A., and Oliveira, F. G. (2024). Enhancing casting manufacturing quality control with vision transformers. In 2024 Brazilian Symposium on Robotics (SBR) and 2024 Workshop on Robotics in Education (WRE), pages 162–167.
Dong, X., Taylor, C. J., and Cootes, T. F. (2018). Small defect detection using convolutional neural network features and random forests. In Proceedings of the European Conference on Computer Vision (ECCV) Workshops, pages 0–0.
Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., Uszkoreit, J., and Houlsby, N. (2021). An image is worth 16x16 words: Transformers for image recognition at scale.
Duan, L., Yang, K., and Ruan, L. (2021). Research on automatic recognition of casting defects based on deep learning. IEEE Access, 9:12209–12216.
Kumaresan, S., Aultrin, K. J., Kumar, S., and Anand, M. D. (2021). Transfer learning with cnn for classification of weld defect. Ieee Access, 9:95097–95108.
Omar, F., Sohrab, H., Saad, M., Hameed, A., and Bakhsh, F. I. (2022). Deep learning binary-classification model for casting products inspection. In 2022 2nd Int. Conf. on Power Electronics IoT Applications in Renewable Energy and its Control (PARC), pages 1–6.
Rocha, C. S., Menezes, M. A., and Oliveira, F. G. (2016). Detecção automática de microcomponentes smt ausentes em placas de circuito impresso. In Workshop on Industry Applications (WIA) in the 29th Conference on Graphics, Patterns and Images (SIBGRAPI 2016), São José dos Campos, Sp, Brazil, volume 1.
Silva., C. N., Ferreira., N. P., Meireles., S. S., Otani., M., J. da Silva., V., O. de Freitas., C. A., and Oliveira., F. G. (2022). The visual inspection of solder balls in semiconductor encapsulation. In Proceedings of the 19th International Conference on Informatics in Control, Automation and Robotics - ICINCO, pages 750–757. INSTICC, SciTePress.
Simonyan, K. and Zisserman, A. (2015). Very deep convolutional networks for large-scale image recognition.
Tan, M. and Le, Q. (2019). Efficientnet: Rethinking model scaling for convolutional neural networks. In Int. conference on machine learning, pages 6105–6114. PMLR.
De Souza Gil, E., De Abreu Dias, L. M., De Souza Brito, A., and Oliveira, F. G. (2024). Enhancing casting manufacturing quality control with vision transformers. In 2024 Brazilian Symposium on Robotics (SBR) and 2024 Workshop on Robotics in Education (WRE), pages 162–167.
Dong, X., Taylor, C. J., and Cootes, T. F. (2018). Small defect detection using convolutional neural network features and random forests. In Proceedings of the European Conference on Computer Vision (ECCV) Workshops, pages 0–0.
Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., Uszkoreit, J., and Houlsby, N. (2021). An image is worth 16x16 words: Transformers for image recognition at scale.
Duan, L., Yang, K., and Ruan, L. (2021). Research on automatic recognition of casting defects based on deep learning. IEEE Access, 9:12209–12216.
Kumaresan, S., Aultrin, K. J., Kumar, S., and Anand, M. D. (2021). Transfer learning with cnn for classification of weld defect. Ieee Access, 9:95097–95108.
Omar, F., Sohrab, H., Saad, M., Hameed, A., and Bakhsh, F. I. (2022). Deep learning binary-classification model for casting products inspection. In 2022 2nd Int. Conf. on Power Electronics IoT Applications in Renewable Energy and its Control (PARC), pages 1–6.
Rocha, C. S., Menezes, M. A., and Oliveira, F. G. (2016). Detecção automática de microcomponentes smt ausentes em placas de circuito impresso. In Workshop on Industry Applications (WIA) in the 29th Conference on Graphics, Patterns and Images (SIBGRAPI 2016), São José dos Campos, Sp, Brazil, volume 1.
Silva., C. N., Ferreira., N. P., Meireles., S. S., Otani., M., J. da Silva., V., O. de Freitas., C. A., and Oliveira., F. G. (2022). The visual inspection of solder balls in semiconductor encapsulation. In Proceedings of the 19th International Conference on Informatics in Control, Automation and Robotics - ICINCO, pages 750–757. INSTICC, SciTePress.
Simonyan, K. and Zisserman, A. (2015). Very deep convolutional networks for large-scale image recognition.
Tan, M. and Le, Q. (2019). Efficientnet: Rethinking model scaling for convolutional neural networks. In Int. conference on machine learning, pages 6105–6114. PMLR.
Published
2025-07-01
How to Cite
DIAS, Lucas Matos A.; GIL, Emanuelle S.; BRITO, Alternei S.; OLIVEIRA, Felipe G..
Improving Quality Control in Casting Manufacturing Using Vision Transformers. In: ICET TECHNOLOGY CONFERENCE (CONNECTECH), 2. , 2025, Itacoatiara/AM.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2025
.
p. 262-269.
DOI: https://doi.org/10.5753/connect.2025.12340.
