Article Details

Abstract

Addressing the issues associated with traditional tea plant disease and pest detection relying on manual labor, low efficiency, and strong subjectivity, this paper designs and implements an intelligent tea plant disease detection system using Computer Vision and Machine Learning. The system adopts an end-to-end architecture of image enhancement-region detection-feature extraction-classification recognition. It first applies the Retinex algorithm to enhance the contrast of disease spots' features, then uses R-CNN to accurately locate the disease region, and extracts deep features using the VGGnet. Finally, three classification models namely SVM, XGBoost and Random Forest, are trained and their performance is compared. This system provides technical support for rapid diagnosis and precise prevention and control of tea plantation diseases, and has strong engineering practicality and application potential.

Keywords

tea plant disease and pest detection; Machine Learning; VGG; model comparison; Computer Vision

References

1. [1] Wishing you good luck, Zhang Fengwei Research on Tea Disease Classification Based on Deep Learning [J]. Fujian Tea, 2025, 47 (12): 50-52
2. [2] Lu Hongfei, Party Min, Wang Renjun, etc Comparative study on YOLO based tea disease recognition algorithms [J]. Agricultural Development and Equipment, 2025 (10): 151-153
3. [3] Lu Chengbo Intelligent Monitoring System for Tea Plant Diseases and Pests in Drone Operation Scenarios [D]. Hangzhou: Hangzhou University of Electronic Science and Technology, 2025
4. [4] Li Wenmin Research on Lightweight Detection Method of Tea Pests and Diseases Based on Improved YOLOv8s [D]. Xianyang: Northwest A&F University, 2025
5. [5] BATZIOU E，IOANNIDIS K，PATRAS I，et al. HDD-Unet: A Unet-based Architecture for Low-Light Image Enhancement [J/OL].Image and Vision Computing，2026，167：105889[2025-10-05].https://doi.org/10.1016/j.imavis.2025.105889.
6. [6] Sun Mingyang Research on Low Light Image Enhancement Algorithm Based on Retinex Theory [D]. Hohhot: Inner Mongolia University, 2025
7. [7] Wang Xinxin Research on Low Light Image Enhancement Algorithm Based on Retinex and Deep Learning [D]. Hohhot: Inner Mongolia University, 2025
8. [8] Jiang Ke Research and Implementation of Tomato Disease and Pest Detection Method Based on Improved SVM [D]. Shenyang: Shenyang University of Technology, 2023
9. [9] Meng Zhichao, He Leiying, Du Xiaoqiang, etc Classification of Camellia oleifera varieties based on Enhanced VGG16 [J]. Journal of Agricultural Engineering, 2022, 38 (10): 176-181
10. [10] Yang Zenghui Research on Image Enhancement Algorithm Based on Deep Learning [D]. Nanjing: Nanjing University of Posts and Telecommunications, 2024

Funding

Innovation and Entrepreneurship Training Program for College Students in Jiangsu Province (202512056061Y); Nantong Science and Technology Plan Project (MSZ2025047); Training Program for Young and Middle aged Backbone Teachers at Nantong University of Science and Technology (ZQNGGJS202225); Algorithm Design Competition Innovation Experimental Class (T2025019C)

About the authors

Luo Jiajia (2005-), female, Han ethnicity, from Guiyang, Guizhou, undergraduate studies, research direction: artificial intelligence; Corresponding author: Chen Shuqian (1992-), female, Han ethnicity, from Nantong, Jiangsu, lecturer, master's student, research direction: artificial intelligence.