Design and Development of Smart Farming System for Monitoring and Bird Pest Control Based on Raspberry Pi 4 with Implementation of YOLOv5 Algorithm

This research proposes a novel solution to address the issue of bird pests in agricultural fields by utilizing a Raspberry Pi 4 Model B, a webcam, and a connected speaker. Leveraging the capabilities of the Internet of Things and computer vision technologies, the webcam employs the YOLOv5 algorithm and OpenCV for real-time bird detection. The system incorporates a Telegram chatbot for instant notifications and allows for photo capture when birds cross predefined boundaries. Extensive testing showcases the webcam's ability to detect birds under various conditions. In open spaces, the detection range spans approximately 80-400 cm, while in environments resembling rice fields, the range is around 80-200 cm, although false positives may occur at a distance of 200 cm. The Raspberry Pi 4 Model B demonstrates a processing speed of 0.78-0.8 frames per second. During a 1-hour temperature test, the temperature of the Raspberry Pi 4 Model B ranged from 58.4 to 61.8 °C. Furthermore, the implementation of a speaker emitting loud noises effectively repels birds within the covered area, although some birds may still return to that area.