YOLOV8-CMS: a high-accuracy deep learning model for automated citrus leaf disease classification and grading
Hongyan Zhu1,2, Dani Wang3,4, Yuzhen Wei5
1Guangxi Key Laboratory of Brain-Inspired Computing and Intelligent Chips, School of Electronic and Information Engineering, Guangxi Normal University, Guilin, 541004, China. hyzhu-zju@foxmail.com.
Plant Methods
|June 23, 2025
Summary
This study introduces an advanced deep learning model, YOLOV8-CMS, for accurate citrus leaf disease classification and severity grading. The model significantly enhances precision agriculture by improving disease detection and management in citrus production.
Area of Science:
- Agricultural Science
- Computer Science
- Plant Pathology
Background:
- Citrus leaf diseases pose a significant threat to agricultural productivity and fruit quality.
- Accurate identification and classification of these diseases are crucial for effective management.
Purpose of the Study:
- To develop and evaluate a deep learning-based approach for classifying citrus leaf diseases.
- To implement a segmentation method for quantifying disease severity based on lesion-to-leaf area ratios.
Main Methods:
- A novel classification model, YOLOV8-CMS, integrating CSPPC, MultiDimen attention, and SpatialConv modules was developed.
- A segmentation technique was employed to delineate leaf and lesion areas for severity assessment.
- Performance was benchmarked against various YOLO architectures (YOLOV8 series, YOLOV5, YOLOV3).
Main Results:
- The YOLOV8-CMS model achieved 98.2% mAP50 for healthy vs. diseased leaf classification.
- The model demonstrated 97.9% accuracy in multi-class disease classification.
- Segmentation enabled precise, quantitative disease severity grading.
Conclusions:
- The YOLOV8-CMS model offers superior performance for citrus leaf disease classification compared to traditional methods.
- The integrated segmentation approach provides accurate, quantitative disease severity assessment.
- Deep learning shows strong potential for advancing precision agriculture and citrus disease management.
