High-precision pest and disease detection in greenhouses using the novel IM-AlexNet framework

Area of Science:

• Agricultural Science
• Computer Vision
• Artificial Intelligence

Background:

• Greenhouse vegetable production faces challenges with high pest and disease incidence due to closed environments.
• Accurate and early identification of pests and diseases is crucial for effective management and crop yield.

Purpose of the Study:

• To develop and evaluate an improved deep learning model for enhanced identification of pests and diseases in greenhouse vegetables.
• To assess the model's performance against traditional methods and other advanced models.

Main Methods:

• An improved AlexNet (IM-AlexNet) model was proposed, integrating ReLU6, batch normalization, and GoogleNet Inception-v3.
• The model's performance was evaluated using metrics including Precision, Recall, F1-score, and Mean Average Precision (MAP).
• Comparative analysis was conducted against AlexNet, Convolutional Neural Network (CNN), and YOLO-v7 models.

Main Results:

• The IM-AlexNet model demonstrated superior performance in Precision, Recall, F1, and MAP compared to traditional models.
• Achieved a Mean Average Precision (MAP) of 88.91%, outperforming AlexNet (78.14%), CNN (80.31%), and YOLO-v7 (83.77%) by significant margins.
• Exhibited strong generalization capabilities, reduced missed detections, and improved target recognition in complex backgrounds, especially under small sample conditions.

Conclusions:

• The IM-AlexNet model provides a robust and accurate solution for intelligent pest and disease monitoring in greenhouse settings.
• This technology can significantly aid greenhouse vegetable growers in reducing pesticide application and promoting environmental sustainability.
• The study lays the groundwork for future advancements in AI-driven agricultural pest management systems.