EfficientNet-B۳ for iron ore pellet quality control: real-world image classification with high accuracy

This study proposes a deep learning–based quality control method for iron ore pellet production using real-world image classification. An EfficientNet-B۳ architecture classifies pellet images into four size categories: very small, small, medium, and large. Trained on ۱۷۴۹۲ images captured under realistic conditions, the model achieved a classification accuracy of ۹۹.۹%, outperforming alternative architectures, including ResNet۵۰, VGG۱۶, and MobileNet. Additional performance metrics, such as precision, sensitivity, and Matthews correlation coefficient (MCC) which were further confirmed the robustness of the approach. The results demonstrated the potential of deep learning for automating pellet size monitoring and highlight its industrial relevance for improving efficiency and quality in steel production.

کلیدواژه ها:

Iron Ore Pelletization ، Image classification ، Deep learning ، EfficientNet ، quality control