Real-Time AI-Driven IoT System for Sustainable Wildlife Monitoring and Poaching Detection

AI-driven wildlife surveillance and poaching detection system combine deep learning-based object detection with IoT technologies to support conservation efforts. Powered by a Raspberry Pi 5, it records live video from a Logitech C270 webcam and streams the feed securely to a Flask-based web application for real-time viewing. The YOLOv8 model is trained to detect 18 wildlife species while ignoring humans. When a human is detected, the system underlines the bounding box in red, saves details (timestamp, species count, confidence score), and produces an alert sound to notify conservation teams. Detections are saved to a CSV file for the monitoring of intrusions and wildlife activity. A bird sound classification module processes calls but only executes when video detection is off for detecting performance optimizations. Secure remote access is provided through a Cloudflare tunnel, and ONNX model conversion and a lightweight Flask implementation maximize efficiency. Scalable in design, several Raspberry Pi units can be installed at various locations, creating an AI-based monitoring network. The system aids conservationists by providing automated species tracking, poaching notifications, and behavioral data gathering, providing a cost-effective and smart solution for biodiversity and sensitive ecosystem protection.