Tomato leaf disease detection method based on improved YOLOv8n.

Area of Science:

• Agricultural Science
• Computer Vision
• Plant Pathology

Background:

• Precision agriculture demands accurate, automated detection of tomato leaf diseases.
• Tomato Yellow Virus Leaf poses a significant challenge due to its unique characteristics, often missed by traditional methods.
• Inaccurate disease detection negatively impacts tomato yield, quality, and farmer response times.

Purpose of the Study:

• To develop an optimized object detection model for precise identification of Tomato Yellow Virus Leaf.
• To enhance the accuracy and efficiency of automated disease detection in smart agriculture.
• To improve upon existing image recognition methods for challenging plant diseases.

Main Methods:

• An optimized YOLOv8n algorithm was proposed, integrating a C2f-DynamicConv module for adaptive feature representation.
• The SimAM attention mechanism was incorporated to improve focus on critical disease features and filter irrelevant information.
• Dysample upsampling and GIoU loss function were utilized to refine feature reconstruction and enhance bounding box regression accuracy.

Main Results:

• The optimized model achieved an average precision of 81.8%, precision of 77.1%, and recall of 77.4%.
• Significant improvements were observed in detection accuracy and localization precision compared to existing methods.
• The enhanced model demonstrated superior computational efficiency for tomato leaf disease detection.

Conclusions:

• The proposed optimized YOLOv8n model effectively addresses the challenge of detecting Tomato Yellow Virus Leaf.
• This advancement offers a more accurate and efficient solution for disease identification in precision agriculture.
• The study highlights the potential of advanced deep learning techniques in smart farming applications.