YOLOv8 assisted smart agriculture approach for tea (Camellia sinensis) leaf disease detection

SANJEET KUMAR  BORAH; KANDARPA  KALITA; LAXMI NARAYAN  SETHI; AVINASH  KUMAR

doi:10.56093/ijas.v96i7.174068

Authors

SANJEET KUMAR BORAH Assam Down Town University, Panikhaiti, Guwahati, Assam 781 026, India
KANDARPA KALITA Assam Down Town University, Panikhaiti, Guwahati, Assam 781 026, India
LAXMI NARAYAN SETHI Assam Down Town University, Panikhaiti, Guwahati, Assam 781 026, India
AVINASH KUMAR Assam Down Town University, Panikhaiti, Guwahati, Assam 781 026, India

https://doi.org/10.56093/ijas.v96i7.174068

Keywords:

Computer vision techniques, Deep learning models, Disease detection, Precision agriculture, YOLOv8 architecture

Abstract

Tea (Camellia sinensis L.) is a major agricultural crop whose yield and quality are highly affected by foliar diseases that impair plant health and reduce market value. Many diseases exhibit similar early symptoms, making manual field inspection slow, inconsistent and prone to misdiagnosis under variable field conditions. These limitations create a strong need for accurate, scalable and automated disease detection methods to support effective tea crop management. To overcome these challenges, the present study was carried out during July–October 2025, that presents a YOLOv8-based deep learning framework capable of automatically detecting and localising multiple tea leaf diseases using images captured under natural field conditions. The dataset included seven major disease classes and a healthy leaf category with total 5,276 images, each annotated using normalised YOLO bounding-box formats and enhanced through diverse augmentation strategies to strengthen model generalisation. The YOLOv8-s model was trained for 100 epochs using a composite loss function that combines Complete IoU (CIoU) for accurate bounding- box regression with classification and objectness losses. Model performance was assessed using standard detection metrics. The trained model achieved a mean average precision @0.5 of 97.68%, a mean average Precision @0.5:0.95 of 97.68%, precision of 93.61% and recall of 95.31%, reflecting substantial improvement over the training period. Further diagnostic evaluations including confusion matrices, precision-recall curves and confidence-based performance analyses confirmed the model’s robustness, even for visually similar disease symptoms such as brown blight and tea algal leaf spot. Overall, the proposed YOLOv8 framework demonstrated high accuracy and computational efficiency for tea leaf disease detection. Its strong performance under real field variability highlights its potential for integration into automated plantation monitoring systems, UAV-assisted crop surveillance and mobile-based advisory tools to support precision agriculture.

Downloads

Download data is not yet available.

References

Abbas A, Zhang Z, Zheng H, Alami M M and Alrefaei A F. 2023. Drones in plant disease assessment, efficient monitoring and detection: A way forward to smart agriculture. Agronomy 13(6): 1524. https://doi.org/10.3390/agronomy13061524 DOI: https://doi.org/10.3390/agronomy13061524

Ayon R, Mubin A S, Hridoy Al-A G, Mojumder M U, Chakraborty N and Islam Md J. 2024. Deep learning for tea leaf disease detection: Comparing VGG16, vision transformer, efficientnetb3 and xception. (In) 2024 15th International Conference on Computing Communication and Networking Technologies (ICCCNT), Kamand, 24–28 June, pp. 1–6. https://doi.org/10.1109/ICCCNT61001.2024.10726203 DOI: https://doi.org/10.1109/ICCCNT61001.2024.10726203

Berger A and Guda S. 2020. Threshold optimisation for F measure of macro-averaged precision and recall. Pattern Recognition 102: 107250. https://doi.org/10.1016/j.patcog.2020.107250 DOI: https://doi.org/10.1016/j.patcog.2020.107250

Datta S and Gupta N. 2023. A novel approach for the detection of tea leaf disease using deep neural network. Procedia Computer Science 218: 2273–86. https://doi.org/10.1016/j.procs.2023.01.203 DOI: https://doi.org/10.1016/j.procs.2023.01.203

Ghafar A, Chen C, Atif A S, Ur Rehman Z and Rahman G. 2024. Visualising plant disease distribution and evaluating model performance for deep learning classification with YOLOv8. Pathogens 13(12): 1032. https://doi.org/10.3390/pathogens13121032 DOI: https://doi.org/10.3390/pathogens13121032

Heng Q, Yu S and Zhang Y. 2024. A new AI-based approach for automatic identification of tea leaf disease using deep neural network based on hybrid pooling. Heliyon 10(5): e26465. https://doi.org/10.1016/j.heliyon.2024.e26465 DOI: https://doi.org/10.1016/j.heliyon.2024.e26465

Li Y, Lu Y, Liu H, Bai J, Yang C, Yuan H, Li X and Xiao Q. 2024. Tea leaf disease and insect identification based on improved MobileNetV3. Frontiers in Plant Science 15: 1459292. https://doi.org/10.3389/fpls.2024.1459292 DOI: https://doi.org/10.3389/fpls.2024.1459292

Olorunshola O E, Irhebhude M E and Evwiekpaefe A E. 2023. A comparative study of YOLOv5 and YOLOv7 object detection algorithms. Journal of Computing and Social Informatics 2(1): 1–12. https://doi.org/10.33736/jcsi.5070.2023 DOI: https://doi.org/10.33736/jcsi.5070.2023

Pandey A K, Sinniah G D, Babu A and Tanti A. 2021. How the global tea industry copes with fungal diseases – Challenges and opportunities. Plant Disease 105(7): 1868–79. https://doi. org/10.1094/PDIS-09-20-1945-FE DOI: https://doi.org/10.1094/PDIS-09-20-1945-FE

Panthee D R, Pandey A and Paudel R. 2024. Multiple foliar fungal disease management in tomatoes: A comprehensive approach. DOI: https://doi.org/10.20944/preprints202306.1268.v1

International Journal of Plant Biology 15(1): 69–93. https:// doi.org/10.3390/ijpb15010007

Preanto S A, Ahad M T, Emon Y R, Mustofa S and Alamin Md. 2024. A semantic segmentation approach on sweet orange leaf diseases detection utilising YOLO. arXiv. https://doi.org/10.48550/arXiv.2409.06671

Rahman H, Ahmed J, John P, Salam M A and Rabbi M F. 2024. Automated detection of selected tea leaf diseases in Bangladesh with convolutional neural network. Scientific Reports 14: 14097. https://doi.org/10.1038/s41598-024-62058-3 DOI: https://doi.org/10.1038/s41598-024-62058-3

Ruby E D K, Amrithayogam G, Sasi G, Chitra G, Choubey A and Gopalkrishnan S. 2024. Advanced image processing techniques for automated detection of healthy and infected leaves in agricultural systems. Mesopotamian Journal of Computer Science 4(1): 44–52. https://doi.org/10.58496/MJCSC/2024/006 DOI: https://doi.org/10.58496/MJCSC/2024/006

Singh R, Sharma N, Chauhan R, Choudhary A and Gupta R. 2023. Automatic approach based on deep learning for tea leaf disease detection. (In) 2nd International Conference on Futuristic Technologies (INCOFT), Belagavi, Karnataka, 24–26th November 2023, pp. 1–5. https://doi.org/10.1109/ INCOFT60753.2023.10425660 DOI: https://doi.org/10.1109/INCOFT60753.2023.10425660

Srilatha A and Praveen P. 2025. CNN-based plant leaf disease identification and classification for sustainable agriculture. (In) 6th International Conference on Mobile Computing and Sustainable Informatics (ICMCSI), Goathgaun, Nepal, 7–8th January 2025, pp. 1459–65. https://doi.org/10.1109/ ICMCSI64620.2025.10883522 DOI: https://doi.org/10.1109/ICMCSI64620.2025.10883522

Srivastav S, Guleria K and Sharma S. 2023. Tea leaf disease detection using deep learning-based convolutional neural networks. (In) 2023 IEEE World Conference on Applied Intelligence and Computing (AIC), Varanasi, 29–30 July, pp. 569–74. https://doi.org/10.1109/AIC57670.2023.10263835 DOI: https://doi.org/10.1109/AIC57670.2023.10263835

Sun X, Mu S, Xu Y, Cao Z and Su T. 2018. Image recognition of tea leaf diseases based on convolutional neural network. (In) International Conference on Security, Pattern Analysis and Cybernetics (SPAC), Jinan, China, 14–17 December, pp. 309. https://doi.org/10.1109/SPAC46244.2018.8965555 DOI: https://doi.org/10.1109/SPAC46244.2018.8965555

Thangaraj R, Vadivel B, Arvind T, Chandhru L, Harish K and Eswaran S. 2023. Tea leaf disease identification using improved convolution neural network. (In) International Conference on Digital Applications, Transformation and Economy (ICDATE), Miri, Malaysia, 14–16 July, pp. 1–5. https://doi.org/10.1109/ICDATE58146.2023.10248508 DOI: https://doi.org/10.1109/ICDATE58146.2023.10248508

Upadhyay A, Chandel N S, Singh K P, Chakraborty S K and Nandede B M. 2025. Deep learning and computer vision in plant disease detection: A comprehensive review of techniques, models and trends in precision agriculture. Artificial Intelligence Review 58(3): 92. https://doi.org/10.1007/s10462-024-11100-x DOI: https://doi.org/10.1007/s10462-024-11100-x

Vilar A M, Garcia L, Garcia S A, Asorey C R and Garcia H J. 2024. Enhancing precision agriculture pest control: A generalised deep learning approach with YOLOv8-based insect detection. IEEE Access 12: 84420–34. https://doi.org/10.1109/ACCESS.2024.3413979 DOI: https://doi.org/10.1109/ACCESS.2024.3413979

Wang C, Li H, Deng X, Liu Y and Wu T. 2024. Improved you only look once v.8 model based on deep learning: Precision detection and recognition of fresh leaves from yunnan large- leaf tea tree. Agriculture 14(12): 2324. https://doi.org/10.3390/ agriculture14122324 DOI: https://doi.org/10.3390/agriculture14122324

Zhang J, Guo H, Guo J and Zhang J. 2023. An information entropy masked vision transformer (IEM-ViT) model for recognition of tea diseases. Agronomy 13(4): 1156. https://doi.org/10.3390/ agronomy13041156 DOI: https://doi.org/10.3390/agronomy13041156