VMF-SSD: A Novel V-Space Based Multi-Scale Feature Fusion SSD for Apple Leaf Disease Detection

Area of Science:

• Agricultural Science
• Computer Vision
• Plant Pathology

Background:

• Apple leaf diseases significantly impact fruit quality and yield.
• Accurate disease detection is vital for disease management and agricultural productivity.
• Existing methods struggle with detecting lesions of varying sizes, limiting accuracy.

Purpose of the Study:

• To propose a novel method for accurate apple leaf disease detection.
• To address challenges in detecting diseased spots of different sizes.
• To improve the overall performance and reliability of apple leaf disease identification.

Main Methods:

• Developed VMF-SSD (V-space-based Multi-scale Feature-fusion SSD) for disease detection.
• Employed multi-scale feature extraction to capture varied lesion sizes, especially small ones.
• Integrated a V-space-based location branch to enhance texture features and spot localization.
• Utilized attention mechanisms to weigh feature importance across different scales.

Main Results:

• VMF-SSD achieved a mean Average Precision (mAP) of 83.19%.
• The method demonstrated a detection speed of 27.53 Frames Per Second (FPS).
• Experimental results confirm competitive performance on the apple leaf disease detection task.

Conclusions:

• The VMF-SSD method offers a robust solution for detecting apple leaf diseases with varying lesion sizes.
• The proposed approach meets the practical requirements for agricultural production applications.
• This advancement contributes to healthier apple growth and industry sustainability.