Evaluation and analysis of passive optical network in investigating real-time cell phone detection in restricted zones

Abstract This research addresses the escalating issue of unauthorized mobile phone use in restricted areas such as examination halls, defense installations, and secure business environments. To mitigate risks ranging from academic dishonesty to data breaches, the study proposes a real-time mobile phone detection system leveraging computer vision, machine learning, and IoT. Utilizing Python and the COCO (Common objects in context) pre-trained object detection model, the system accurately identifies mobile devices within a webcam’s field of view. Upon detection, it captures an image and sends an immediate email alert to designated authorities, enhancing security enforcement. The COCO model’s robustness allows for effective detection under diverse conditions, including low light, partial occlusions, and varied phone types. This eliminates the need for large training datasets, making the system easier and quicker to deploy. Python’s integration capabilities and rich libraries ensure seamless operation and high computational efficiency. Experimental results demonstrate high accuracy and minimal latency, supporting timely responses in sensitive settings. Beyond its immediate applications, this work exemplifies the potential of integrating AI, computer vision, and IoT to tackle real-world challenges. The system’s adaptability and interdisciplinary foundation offer a forward-looking solution for maintaining security in dynamically changing environments, marking a significant advancement in digital surveillance and control mechanisms.