Patrol: An Intelligent Surveillance System for Mobile-Free Environments

The widespread and unauthorized use of mobile phones in no-mobile zones such as examination halls, government offices, and hospitals presents significant security, confidentiality, and productivity challenges. Traditional signal-based detection methods fail to detect phones in airplane mode or disconnected from networks, leaving critical gaps in enforcement. To address this issue, we propose "Phone Patrol," a deep learning-based system designed to detect and report unauthorized mobile phone usage. Phone Patrol uses You Only Look Once version 5 (YOLOv5) for detecting individuals with mobile phones and You Only Look Once version 9 (YOLOv9) for face recognition, enabling identification of violators. Deployed on a Raspberry Pi 5, the system processes real-time video feeds and sends automated SMS alerts to offenders, offering a compact, standalone Internet of Things (IoT) solution. Experimental results demonstrate that YOLOv5 achieved a mean Average Precision at 50% Intersection over Union (mAP50) score of 0.992, showcasing its accuracy in detecting mobile phones. Although the current prototype has limitations in scalability and optimization for larger environments, it presents a robust proof-of-concept for image-based detection independent of network signals, providing a more comprehensive solution to enforce no-mobile zones effectively.