Classification and localization of maize leaf spot disease based on weakly supervised learning.

Area of Science:

• Agricultural Science
• Computer Science
• Artificial Intelligence

Background:

• Accurate crop disease detection is vital for effective crop production monitoring and targeted plant protection.
• Current methods for disease identification and localization can be labor-intensive and may lack precision.

Purpose of the Study:

• To develop and evaluate an AI framework for classifying and localizing maize leaf diseases.
• To improve the interpretability and efficiency of crop disease detection systems.

Main Methods:

• Constructed a dataset of six types of field maize leaf images.
• Integrated lightweight convolutional neural networks (CNNs) with interpretable AI algorithms.
• Employed weakly supervised semantic segmentation using class activation mapping (CAM) for disease spot localization.

Main Results:

• Achieved high classification accuracy and fast detection speeds for maize leaf diseases.
• Demonstrated a mean Intersection over Union (mIoU) of up to 55.302% for disease spot localization.
• Validated the feasibility of using CAM-based weakly supervised learning for crop disease detection.

Conclusions:

• The developed framework successfully combines deep learning with visualization techniques for interpretable and accurate maize disease detection.
• Weakly supervised learning enables precise localization of infected areas on maize leaves.
• The framework supports smart crop disease monitoring and automated plant protection via various devices.