Citrus diseases detection using innovative deep learning approach and Hybrid Meta-Heuristic
Nouman Butt1, Muhammad Munwar Iqbal1, Shabana Ramzan2
1Department of Computer Science, University of Engineering and Technology, Taxila, Pakistan.
Plos One
|January 22, 2025
Summary
This study introduces an automated system for citrus disease classification using deep learning, achieving 99.6% accuracy. This AI-driven approach enhances disease detection in citrus farming for improved crop yield and sustainability.
Area of Science:
- Agricultural Science
- Computer Science
- Plant Pathology
Background:
- Citrus farming is vital to Pakistan's economy, contributing 30% of fruit production, primarily in Punjab.
- Citrus crops face significant disease threats (canker, scab, black spot), reducing fruit quality and yield.
- Manual disease diagnosis is slow, inaccurate, costly, and requires expert knowledge.
Purpose of the Study:
- To develop an automated disease classification system for citrus crops.
- To enhance diagnostic accuracy, efficiency, and cost-effectiveness in disease detection.
- To provide a scalable solution for sustainable citrus farming practices.
Main Methods:
- Implementation of a deep learning framework with optimal feature selection.
- Utilized data augmentation and transfer learning techniques.
- Employed pre-trained models including DenseNet-201 and AlexNet for classification.
Main Results:
- Achieved a classification accuracy of 99.6% on a citrus leaves dataset.
- The proposed automated system demonstrated superior performance compared to existing methods.
- The framework offers a robust and efficient solution for citrus disease identification.
Conclusions:
- The developed deep learning system provides an accurate and efficient method for citrus disease classification.
- This technology supports sustainable agriculture by improving disease management in citrus farming.
- The automated system addresses the limitations of traditional manual diagnosis, offering a scalable solution.
