Peacocks have become a significant threat to agricultural fields, causing crop damage and affecting irrigation infrastructure. Traditional deterrence tools, such as scarecrows and nets, are often inefficient, labor- intensive, and environmentally unsustainable in the long term. To address this problem, this study proposes an automatic peacock deterrent system that combines deep learning-based detection with integrated actuation control. The system uses the YOLOv8 object detection model to detect peacocks in real-time video streams recorded using a web camera. When detected, a dynamically controlled Arduino-operated dual-servo pan–tilt mechanism is used to point a low-power laser at the target and start predator sounds to frighten birds. This approach provides a non-lethal, effective, and automatic visual–auditory deterrence mechanism. The system achieved a detection accuracy of 92.5% and a repellent success rate of approximately 90%, demonstrating effective and real-time performance of the proposed system. The proposed solution is effective in reducing manual intervention, improving crop protection, and supporting sustainable agricultural practices. This study highlights the capabilities of AI-based systems in precision agriculture and the alleviation of human–wildlife conflict.
Balaji et al. (Wed,) studied this question.