EfficientNet-B3 for iron ore pellet quality control: real-world image classification with high accuracy

Articles in Press

Document Type : Research Paper

Authors

Faculty of Electrical and Computer Engineering, Sirjan University of Technology, Sirjan, Iran.

10.22059/ijmge.2025.395689.595261

Abstract

This study proposes a deep learning–based quality control method for iron ore pellet production using real-world image classification. An EfficientNet-B3 architecture classifies pellet images into four size categories: very small, small, medium, and large. Trained on 17492 images captured under realistic conditions, the model achieved a classification accuracy of 99.9%, outperforming alternative architectures, including ResNet50, VGG16, and MobileNet. Additional performance metrics—precision, sensitivity, and Matthews correlation coefficient (MCC)—further confirm the robustness of the approach. The results demonstrate the potential of deep learning for automating pellet size monitoring and highlight its industrial relevance for improving efficiency and quality in steel production.

Keywords

Main Subjects

References

[1] Duan, J., Liu, X., Wu, X., & Mao, C., (2020) Detection and segmentation of iron ore green pellets in images using lightweight U-Net deep learning network. Neural Comput Appl, 32, 5775-5790.
[2] Deo, AJ., Sahoo, A., Behera, SK., & Das, DP., (2022) Vision-based size classification of iron ore pellets using ensembled convolutional neural network. Neural Comput Appl , 34, 18629-18641.
[3] Liu, Y., Wang, X., Zhang, Z., & Deng, F., (2023) A review of deep learning in image classification for mineral exploration. Miner Eng, 204, 108433.
[4] Liao, C., & Tarng, Y., (2009) On-line automatic optical inspection system for coarse particle size distribution. Powder Technol , 189, 508-513.
[5] Budzan, S., & Pawelczyk, M., (2018) Grain size determination and classification using adaptive image segmentation with shape-context information for indirect mill faults detection. In: Advances in Technical Diagnostics. Proc 6th Int Congr Tech Diagn, 6, 215-224.
[6] Ma, W.,Wang, L., Jiang, T., Yang, A., & Zhang, Y., (2023) Overlappingpelletsizedetectionmethod based on marker watershed and GMM image segmentation. Metals, 13, 327.
[7] Subramanyam, V., Patra, P., & Singh, MK., (2017) Automatic image processing-based size char- acterization of green pellets. Int J Autom Smart Technol, 7, 85-91.
[8] Kamilaris, A., & Prenafeta-Boldu, FX., (2018) Deep learning in agriculture: A survey. Comput Electron Agric, 147, 70-90.
[9] Russakovsky, O., Deng, J., Su, H., Krause, J., Satheesh, S., Ma, S., Huang, Z., Karpathy, A., Khosla, A., Bernstein, M. & Berg, A.C., (2015) ImageNet large scale visual recognition challenge. Int J Comput Vis, 115, 211-252.
[10] He, K., Zhang, X., Ren, S., & Sun, J., (2016) Deep residual learning for image recognition. In: Proc IEEE Conf Comput Vis Pattern Recognit, 770-778.
[11] Tan, M., (2019) EfficientNet: Rethinking model scaling for convolutional neural networks. arXiv.
[12] Salehi, M., & Shamsi-Koshki, A., (2024) Classification of iron ore pellets based on size using deep learning. In: Proc 13th Iranian and 3rd Int Conf Comput Vis Image Process. [In Persian]
[13] Weiss, K., Khoshgoftaar, TM., & Wang, D., (2016) A survey of transfer learning. J Big Data, 3, 1-40.
[14] Ashtiani, S.H.M., Javanmardi, S., Jahanbanifard, M., Martynenko, A., & Ver- beek, F.J., (2021). Detection of mulberry ripeness stages using deep learning models. IEEE Access, 9, 100380-100394.
International Journal of Mining and Geo-Engineering

Articles in Press, Accepted Manuscript
Available Online from 01 November 2025

Files

Share

How to cite

Statistics