YOLO-LeafNet: a robust deep learning framework for multispecies plant disease detection with data augmentation.

Area of Science:

• Agricultural Science
• Computer Vision
• Machine Learning

Background:

• Plant diseases cause significant global economic losses.
• Accurate and timely plant disease diagnosis is crucial for mitigating crop damage.
• Existing diagnostic methods may lack efficiency and accuracy.

Purpose of the Study:

• To propose a novel YOLO-LeafNet approach for detecting plant diseases from leaf images.
• To evaluate the performance of YOLO-LeafNet against established models like YOLOv5 and YOLOv8.
• To enhance crop yield and reduce economic impact through improved disease detection.

Main Methods:

• Acquired 8850 leaf images from five public datasets for grape, bell pepper, corn, and potato.
• Applied four image pre-processing operations and five augmentation operations to enhance the dataset.
• Trained and evaluated YOLOv5, YOLOv8, and the proposed YOLO-LeafNet models using precision, recall, and Mean Average Precision (mAP).

Main Results:

• YOLO-LeafNet achieved a precision of 0.985, recall of 0.980, mAP50 of 0.990, and mAP50-95 of 0.940.
• YOLOv8 attained precision of 0.977, recall of 0.975, mAP50 of 0.984, and mAP50-95 of 0.915.
• YOLOv5 achieved precision of 0.861, recall of 0.868, mAP50 of 0.944, and mAP50-95 of 0.815.
• YOLO-LeafNet demonstrated superior performance across all evaluated metrics compared to YOLOv5 and YOLOv8.

Conclusions:

• The proposed YOLO-LeafNet model significantly outperforms YOLOv5 and YOLOv8 in plant disease detection.
• YOLO-LeafNet offers a highly accurate and efficient solution for automated plant disease diagnosis.
• This technology has the potential to substantially reduce crop losses and improve agricultural productivity.