Image Classification in Coal Production Using Deep Neural Networks: A Comprehensive Benchmarking Study

Wenmi Chai; Zhiyao Yang; Rui Zhao; Qian Xiang; Xinxin Niu; Ling Liang

doi:10.62177/jaet.v2i4.958

Authors

Wenmi Chai New Energy Technology Research Institute Co., Ltd., CHN ENERGY Investment Group Co., Ltd. https://orcid.org/0009-0004-9641-5215
Zhiyao Yang National Institute of Clean-and-Low-Carbon Energy
Rui Zhao Chengdu Jinjiang Center for Disease Control and Prevention
Qian Xiang Laboratory of Intelligent Control, PLA Rocket Force University of Engineering https://orcid.org/0000-0001-6810-8446
Xinxin Niu New Energy Technology Research Institute Co., Ltd., CHN ENERGY Investment Group Co., Ltd.
Ling Liang New Energy Technology Research Institute Co., Ltd., CHN ENERGY Investment Group Co., Ltd.

DOI:

https://doi.org/10.62177/jaet.v2i4.958

Keywords:

Coal Separation, Image Classification, Deep Learning, Convolutional Neural Network, Model Benchmarking

Abstract

During the intelligent transformation of the coal industry, image classification technology plays a crucial role in process monitoring, quality inspection, and safety early warning. Taking the DsCGF-1 dataset in the coal production environment as the research object, this study conducts a multi-dimensional performance and efficiency evaluation on 12 mainstream deep learning models, aiming to establish industrial-level model selection criteria for intelligent coal separation. The results indicate that RepVGG-B3 exhibits the optimal comprehensive performance, with an test accuracy of 97.92%, a coal recall rate of 99.8%, and the best AUC value across all categories. Furthermore, RepViT-M3 achieves an test accuracy of 97.85% with a parameter count of merely 9.66M, demonstrating excellent lightweight characteristics, which makes it suitable for resource-constrained scenarios such as underground edge computing. This study establishes a model selection benchmark for coal separation, providing technical support for the development of intelligent sorting systems in industrial scenarios.

Downloads

Download data is not yet available.

References

Kang H, Xie H, Ren S, et al. (2022) Development Strategy of China’s Coal Industry under the Reconstruction of Global Industrial Chain and Energy Supply Chain. Chinese Journal of Engineering Science 24: 26. DOI: https://doi.org/10.15302/J-SSCAE-2022.07.009

National Bureau of Statistics of China. Statistical Communique on National Economic and Social Development of the People's Republic of China in 2024 [EB/OL]. (2025-02-28). https://www.stats.gov.cn/sj/zxfb/202502/t20250228_19588-

html

Lv Z, Fan Y, Sha T, et al. (2025) A large-scale open image dataset for deep learning-enabled intelligent sorting and analyzing of raw coal. Sci Data 12: 403. DOI: https://doi.org/10.1038/s41597-025-04719-0

Liu Y, Zhang Z, Liu X, et al. (2021) Deep Learning Based Mineral Image Classification Combined With Visual Attention Mechanism. IEEE Access 9: 98091–98109. DOI: https://doi.org/10.1109/ACCESS.2021.3095368

Patel AK, Chatterjee S, Gorai AK (2019) Development of a machine vision system using the support vector machine regression (SVR) algorithm for the online prediction of iron ore grades. Earth Sci Inform 12: 197–210. DOI: https://doi.org/10.1007/s12145-018-0370-6

Wang W, Lv Z, Lu H (2021) Research on methods to differentiate coal and gangue using image processing and a support vector machine. International Journal of Coal Preparation and Utilization 41: 603–616. DOI: https://doi.org/10.1080/19392699.2018.1496912

Zhang L, Sui Y, Wang H, et al. (2022) Image feature extraction and recognition model construction of coal and gangue based on image processing technology. Sci Rep 12: 20983. DOI: https://doi.org/10.1038/s41598-022-25496-5

Pu Y, Apel DB, Szmigiel A, et al. (2019) Image Recognition of Coal and Coal Gangue Using a Convolutional Neural Network and Transfer Learning. Energies 12: 1735. DOI: https://doi.org/10.3390/en12091735

Si L, Xiong X, Wang Z, et al. (2020) A Deep Convolutional Neural Network Model for Intelligent Discrimination between Coal and Rocks in Coal Mining Face. Mathematical Problems in Engineering 2020: 1–12. DOI: https://doi.org/10.1155/2020/2616510

Liu Q, Li J, Li Y, et al. (2021) Recognition Methods for Coal and Coal Gangue Based on Deep Learning. IEEE Access 9: 77599–77610. DOI: https://doi.org/10.1109/ACCESS.2021.3081442

Liu H, Xu K (2023) Recognition of gangues from color images using convolutional neural networks with attention mechanism. Measurement 206: 112273. DOI: https://doi.org/10.1016/j.measurement.2022.112273

Cao Z, Fang L, Li R, et al. (2023) Research on image classification of coal and gangue based on a lightweight convolution neural network. Energy Science & Engineering 11: 3042–3054. DOI: https://doi.org/10.1002/ese3.1501

Pengcheng Y, Heng Z, Xuyue K, et al. (2024) Lightweight detection method of coal gangue based on multispectral and improved YOLOv5s. International Journal of Coal Preparation and Utilization 44: 399–414. DOI: https://doi.org/10.1080/19392699.2023.2193739

He K, Zhang X, Ren S, et al. (2016) Deep Residual Learning for Image Recognition, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA, IEEE, 770–778. DOI: https://doi.org/10.1109/CVPR.2016.90

Chollet F (2017) Xception: Deep Learning with Depthwise Separable Convolutions, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI, IEEE, 1800–1807. DOI: https://doi.org/10.1109/CVPR.2017.195

Xie S, Girshick R, Dollár P, et al. (2017) Aggregated Residual Transformations for Deep Neural Networks. DOI: https://doi.org/10.1109/CVPR.2017.634

Tan M, Chen B, Pang R, et al. (2019) MnasNet: Platform-Aware Neural Architecture Search for Mobile. DOI: https://doi.org/10.1109/CVPR.2019.00293

Ding X, Zhang X, Ma N, et al. (2021) RepVGG: Making VGG-style ConvNets Great Again, 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Nashville, TN, USA, IEEE, 13728–13737. DOI: https://doi.org/10.1109/CVPR46437.2021.01352

Tang Y, Han K, Guo J, et al. (2022) GhostNetV2: Enhance Cheap Operation with Long-Range Attention.

Vasu PKA, Gabriel J, Zhu J, et al. (2023) MobileOne: An Improved One millisecond Mobile Backbone. DOI: https://doi.org/10.1109/CVPR52729.2023.00764

Vasu PKA, Gabriel J, Zhu J, et al. (2023) FastViT: A Fast Hybrid Vision Transformer using Structural Reparameterization.

Woo S, Debnath S, Hu R, et al. (2023) ConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoders. DOI: https://doi.org/10.1109/CVPR52729.2023.01548

Wang A, Chen H, Lin Z, et al. (2024) RepViT: Revisiting Mobile CNN From ViT Perspective. DOI: https://doi.org/10.1109/CVPR52733.2024.01506

Qin D, Leichner C, Delakis M, et al. (2024) MobileNetV4 -- Universal Models for the Mobile Ecosystem. DOI: https://doi.org/10.1007/978-3-031-73661-2_5

Yu W, Zhou P, Yan S, et al. (2025) InceptionNeXt: When Inception Meets ConvNeXt. DOI: https://doi.org/10.1109/CVPR52733.2024.00542

Xiang Q, Wang X, Lei L, et al. (2025) Dynamic bound adaptive gradient methods with belief in observed gradients. Pattern Recognition 168: 111819. DOI: https://doi.org/10.1016/j.patcog.2025.111819