AI-Enabled Crop Management Framework for Pest Detection Using Visual Sensor Data.

Area of Science:

• Agricultural Science
• Computer Vision
• Robotics

Background:

• Crop diseases and pest infestations pose significant threats to agricultural productivity.
• Traditional pest detection methods are often labor-intensive or computationally expensive.
• Unmanned Aerial Vehicle (UAV) technology offers potential for efficient agricultural monitoring.

Purpose of the Study:

• To develop and evaluate an optimized deep learning model for UAV-based agricultural pest detection and classification.
• To improve the accuracy and efficiency of pest identification compared to existing methods.
• To contribute to sustainable agriculture through advanced pest management strategies.

Main Methods:

• Modification of the YOLOv5s model with advanced attention modules, expanded CSP modules, and refined feature extraction.
• Utilizing UAVs for aerial data acquisition in agricultural settings.
• Training and testing the model on a dataset comprising five common agricultural pests.

Main Results:

• The proposed optimized YOLOv5s model demonstrated superior performance in pest detection and classification.
• Achieved an average precision of 96.0%, average recall of 93.0%, and mean average precision (mAP) of 95.0%.
• Outperformed various standard YOLOv5 model versions in experimental tests.

Conclusions:

• The optimized YOLOv5s model provides a precise and efficient solution for real-time pest detection using UAVs.
• This technology has significant potential to enhance agricultural production and disease prevention within drone-centric ecosystems.
• The study highlights the effectiveness of integrating advanced deep learning with UAVs for sustainable pest management.