AI-Based Animal Repellent System for Agricultural Fields Using YOLOv5 and OpenCV for Farmer Safeguard

Crop damage from wild and domestic animal intrusions causes significant economic losses, estimated at 20-30% annually in some agricultural regions. This paper proposes a real-time animal re- pellent system utilizing the YOLOv5 deep learning model for precise object detection, integrated with a Flask web application for farm surveillance. The system processes live video feeds, detects intrusions, and triggers deterrents like alarms or flashing lights. Designed to minimize human intervention, it offers scalability, cost-effectiveness, and adaptability to di- verse environments. With multi-camera support, cloud integration, and robust performance in low-light condi- tions, the system achieves 89% precision and 20-30 FPS on standard hardware, advancing smart agriculture by reducing crop losses and labor costs.