Abstract
Accurate grading of flue-cured tobacco leaves is crucial for tobacco quality control and industrial applications. However, traditional manual grading is subjective, inefficient, and labor-intensive, while existing deep learning-based methods often fail to capture multi-scale complementary features, leading to limited grading precision. To address these issues, this paper proposes a novel deep learning framework that integrates multiple pre-trained architectures with hierarchical feature fusion for robust tobacco leaf classification. Specifically, the framework leverages three complementary backbone networks, incorporates a convolutional block attention module to enhance feature discriminability via channel and spatial attention mechanisms, and designs a hierarchical feature fusion module with learnable attention weights to adaptively combine low-level, mid-level, and high-level features. Experimental results confirm that the proposed method achieves superior performance in flue-cured tobacco leaf grading, boasting a remarkable accuracy of 99.95% and effectively capturing the subtle visual characteristics essential for tobacco leaf quality assessment. In conclusion, the proposed architecture provides a comprehensive and reliable solution for flue-cured tobacco leaf grading, with potential applications extended to other fine-grained visual recognition tasks in agricultural product classification.
Similar content being viewed by others
Data availability
The complete source code, sample data sets, and models used in this study are made publicly available on GitHub at: https://github.com/RickPang/tobaccoVision.
References
Wu, Y. et al. TobaccoNet: A deep learning approach for tobacco leaves maturity identification. Expert Syste. Appl. 255, 124675. https://doi.org/10.1016/j.eswa.2024.124675 (2024).
Chen, Y., Lian, C., Chen, X., Gong, T., Peng, X., Chen, F.: Fine-Grained Visual Classification of Flue-Cured Tobacco Leaf Grades Based on Vein Information. In 2024 10th International Conference on Systems and Informatics (ICSAI), pp. 1–6 (2024). https://doi.org/10.1109/ICSAI65059.2024.10893752
Zhao, P. et al. The Bayesian mixture expert recognition model for tobacco leaf curing stages based on feature fusion. Plant Methods 21(1), 86. https://doi.org/10.1186/s13007-025-01384-7 (2025).
Zhang, M. et al. Integrated volatilomic profiles and chemometrics provide new insights into the aroma differences of volatile compounds in filler tobacco leaves of six grades. Ind. Crops Prod. 232, 121236. https://doi.org/10.1016/j.indcrop.2025.121236 (2025).
He, C. et al. Fermentation-driven microbial and metabolic shifts in filler tobacco leaves of different grades. Front. Microbiol. https://doi.org/10.3389/fmicb.2025.1651289 (2025).
Jin, X., Yi, K. & Xu, J. MoADNet: Mobile asymmetric dual-stream networks for real-time and lightweight RGB-D salient object detection. IEEE Trans. Circuits Syst. Video Technol. 32(11), 7632–7645. https://doi.org/10.1109/TCSVT.2022.3180274 (2022).
Jin, X., Jing, P., Wu, J., Xu, J. & Su, Y. Visual sentiment classification via low-rank regularization and label relaxation. IEEE Trans. Cogn. Dev. Syst. 14(4), 1678–1690. https://doi.org/10.1109/TCDS.2021.3135948 (2022) (Accessed 2026-02-09).
Jin, X., Yu, W. & Shi, W. Image manipulation localization via dynamic cross-modality fusion and progressive integration. Neurocomputing 610, 128607. https://doi.org/10.1016/j.neucom.2024.128607 (2024).
Jin, X., Yu, W., Chen, D.-W. & Shi, W. DFD-NAS: General deepfake detection via efficient neural architecture search. Neurocomputing 619, 129129. https://doi.org/10.1016/j.neucom.2024.129129 (2025).
Jin, X. et al. FCMNet: Frequency-aware cross-modality attention networks for RGB-D salient object detection. Neurocomputing 491, 414–425. https://doi.org/10.1016/j.neucom.2022.04.015 (2022).
Li, Q.-L. et al. Minireview on tobacco classification technologies: A vital bridge linking raw leaf properties with end product quality. J. Anal. Appl. Pyrolysis 193, 107398. https://doi.org/10.1016/j.jaap.2025.107398 (2026).
Niu, Q. et al. Tobacco shred varieties classification using Multi-Scale-X-ResNet network and machine vision. Front. Plant Sci. 13, 962664. https://doi.org/10.3389/fpls.2022.962664 (2022).
Chen, Y., Xia, R., Yang, K. & Zou, K. Dual degradation image inpainting method via adaptive feature fusion and U-net network. Appl. Soft Comput. 174, 113010. https://doi.org/10.1016/j.asoc.2025.113010 (2025).
Zhang, J., Yang, J., Qin, Y., Xiao, Z. & Wang, J. MGNet: RGBT tracking via cross-modality cross-region mutual guidance. Neural Netw. 190, 107707. https://doi.org/10.1016/j.neunet.2025.107707 (2025).
Zhang, J., Zhang, S., Li, D., Wang, J. & Wang, J. Crack segmentation network via difference convolution-based encoder and hybrid CNN-Mamba multi-scale attention. Pattern Recogn. 167, 111723. https://doi.org/10.1016/j.patcog.2025.111723 (2025).
Zhang, F. & Zhang, X. Classification and quality evaluation of tobacco leaves based on image processing and fuzzy comprehensive evaluation. Sensors 11(3), 2369–2384. https://doi.org/10.3390/s110302369 (2011).
Dasari, S.K., Chintada, K.R., Patruni, M.: Flue-Cured Tobacco Leaves Classification: A Generalized Approach Using Deep Convolutional Neural Networks. In Cognitive Science and Artificial Intelligence, pp. 13–21. (Springer, 2018). https://doi.org/10.1007/978-981-10-6698-6_2
Zhang, Y. et al. In-field tobacco leaf maturity detection with an enhanced MobileNetV1: Incorporating a feature pyramid network and attention mechanism. Sensors 23(13), 5964. https://doi.org/10.3390/s23135964 (2023).
Xin, X. et al. Intelligent large-scale flue-cured tobacco grading based on deep densely convolutional network. Sci. Rep. 13(1), 11119. https://doi.org/10.1038/s41598-023-38334-z. (2023) (Accessed 2026-02-09).
Li, Q., Lin, H., Hu, J., Wang, H.: Automated Tobacco Leaf Grading System Based on Deep Learning. In 2023 IEEE 3rd International Conference on Software Engineering and Artificial Intelligence (SEAI), pp. 40–46 (2023). https://doi.org/10.1109/SEAI59139.2023.10217575
Wei, X. et al. Classification method for folded flue-cured tobacco based on hyperspectral imaging and conventional neural networks. Ind. Crops Prod. 212, 118279. https://doi.org/10.1016/j.indcrop.2024.118279 (2024).
Wei, Y., Usman, M., Bilal, H.: InspectionV3: Enhancing Tobacco Quality Assessment with Deep Convolutional Neural Networks for Automated Workshop Management. arXiv (2025). https://doi.org/10.48550/arXiv.2505.16485
Chen, Y., Chen, L., Xia, R., Yang, K. & Zou, K. CAAT: Image super-resolution algorithm via channel attention and transformer. Array 28, 100628. https://doi.org/10.1016/j.array.2025.100628 (2025).
Xiong, H. et al. DiffuCNN: Tobacco disease identification and grading model in low-resolution complex agricultural scenes. Agriculture 14(2), 318. https://doi.org/10.3390/agriculture14020318 (2024).
Zhu, H. et al. Prediction of typical gas components in cigarette smoke based on transformer. Eng. Res. Express 7(1), 015408. https://doi.org/10.1088/2631-8695/ad81cb (2025).
Zhao, P. et al. TCSRNet: A lightweight tobacco leaf curing stage recognition network model. Front. Plant Sci. https://doi.org/10.3389/fpls.2024.1474731 (2024).
Chen, Y., Bin, J. & Kang, C. Application of machine vision and convolutional neural networks in discriminating tobacco leaf maturity on mobile devices. Smart Agric. Technol. 5, 100322. https://doi.org/10.1016/j.atech.2023.100322 (2023).
Bell, J., Dee, H.M.: Leaf Segmentation through the Classification of Edges. arXiv (2019). https://doi.org/10.48550/arXiv.1904.03124
Jianqiang, Z., Panpan, Y., Weijuan, L., Yanmei, Y., Tianjun, Y., Ying, H., Changyu, L.: Rapid and Automatic Classification of Tobacco Leaves Using a Hand-Held DLP-based NIR Spectroscopy Device. Journal of the Brazilian Chemical Society (2019) https://doi.org/10.21577/0103-5053.20190105
Hong, L. et al. Identifying the geographical origin of tobacco leaf by strontium and lead isotopic with mineral elemental fingerprint. Int. J. Chem. Eng. 2022, 1–10. https://doi.org/10.1155/2022/5949770 (2022).
Wang, D. & Yang, S. X. Broad learning system with takagi-sugeno fuzzy subsystem for tobacco origin identification based on near infrared spectroscopy. Appl. Soft Comput. 134, 109970. https://doi.org/10.1016/j.asoc.2022.109970 (2023).
Liu, H., Tian, L., Wang, L., Zhang, Z., Li, J., Liu, X., zheng, B., Ma, H., Wang, Y., Li, J.: Real-Time Grading of Roasted Tobacco Using near Infrared Spectroscopy Technology. Microchemical Journal 204, 110963 (2024) https://doi.org/10.1016/j.microc.2024.110963
Wu, X. et al. Tobacco leaves maturity classification based on deep learning and proximal hyperspectral imaging. Anal. Lett. 57(13), 2034–2049. https://doi.org/10.1080/00032719.2023.2284834 (2024).
Chen, D., Feng, L., Sun, H., Zhong, R., Wang, C., Zhang, X., Zhang, K., Bu, L.-d., Yang, W., Liu, K., Chen, H., Wang, S.: Analysis of Differences in Cigar Tobacco Leaves from Different Regions Based on GC-IMS and LC-MS Metabolomics Techniques. Frontiers in Plant Science 16 (2025) https://doi.org/10.3389/fpls.2025.1557190
Zhang, X., Liu, J., Wang, L., Li, Q., Xu, Z., Ren, Z., He, Q.: Automated Tobacco Leaf Grading Using Visible and Near-Infrared Spectral Images. In Li, W., Wang, H. (eds.) Third International Symposium on Computer Applications and Information Systems (ISCAIS 2024), p. 97. SPIE, Wuhan, China (2024). https://doi.org/10.1117/12.3034935
Chen, H. et al. Classification models for tobacco mosaic virus and potato virus y using hyperspectral and machine learning techniques. Front. Plant Sci. https://doi.org/10.3389/fpls.2023.1211617 (2023).
Don Mariano Marcos Memorial State University, Philippines, Marzan, C.S., Ruiz Jr., C.R.: Automated Tobacco Grading Using Image Processing Techniques and a Convolutional Neural Network. Int. J. Mach. Learn. Comput. 9(6), 807–813 (2019) https://doi.org/10.18178/ijmlc.2019.9.6.877
Liu, Z., Zhang, Q., Wang, P., Li, Z., Wang, H.: Automated Classification of Stems and Leaves of Potted Plants Based on Point Cloud Data. Biosystems Engineering 200, 215–230 (2020) https://doi.org/10.1016/j.biosystemseng.2020.10.006
Liao, Y.-H., Zhang, S.: Combining Multispectral and High-Resolution 3D Imaging for Leaf Vein Segmentation and Density Measurement. Frontiers in Plant Science 16 (2025) https://doi.org/10.3389/fpls.2025.1560220
Xie, K. et al. Automated 3D Segmentation of Plant Organs via the Plant-MAE: A Self-Supervised Learning Framework. Plant Phenomics 7(2), 100049. https://doi.org/10.1016/j.plaphe.2025.100049 (2025).
Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., Uszkoreit, J., Houlsby, N.: An Image Is Worth 16x16 Words: Transformers for Image Recognition at Scale. arXiv (2021). https://doi.org/10.48550/arXiv.2010.11929
Touvron, H., Cord, M., Douze, M., Massa, F., Sablayrolles, A., Jegou, H.: Training data-efficient image transformers & distillation through attention. In: Meila, M., Zhang, T. (eds.) Proceedings of the 38th International Conference on Machine Learning. Proceedings of Machine Learning Research, vol. 139, pp. 10347–10357. PMLR, Virtual Event (2021). https://proceedings.mlr.press/v139/touvron21a.html
Liu, Z., Lin, Y., Cao, Y., Hu, H., Wei, Y., Zhang, Z., Lin, S., Guo, B.: Swin Transformer: Hierarchical Vision Transformer Using Shifted Windows. In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pp. 10012–10022. IEEE, Virtual Event (2021)
Touvron, H., Cord, M., Sablayrolles, A., Synnaeve, G., Jégou, H.: Going Deeper With Image Transformers. In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pp. 32–42. IEEE, Virtual Event (2021)
Acknowledgements
The authors would like to thank all employees of China Tobacco Shandong Industrial Co., Ltd. for their support and cooperation to the project.
Funding
This work is supported by the Scientific Research Foundation for Ph.D (No.WFU2023BS47).
Author information
Authors and Affiliations
Contributions
Shunpeng Pang and Xiaowei Xin contributed equally to this work. Shunpeng Pang: Conceptualization, methodology, formal analysis, investigation, resources, writing—original draft preparation, writing—review and editing, visualization. Xiaowei Xin: Conceptualization, methodology, formal analysis, investigation. Wei Ge:Validation, data curation. Yonghui Zhang: Validation, writing—review and editing. Junhua Jia: Conceptualization, formal analysis, investigation, writing—original draft preparation, visualization.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Pang, S., Xin, X., Ge, W. et al. High-precision automated grading of flue-cured tobacco leaves based on hierarchical feature fusion. Sci Rep (2026). https://doi.org/10.1038/s41598-026-45252-3
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-026-45252-3
