Development of a real-time security management system for restricted access areas using computer vision and deep learning

Abstract

The safety of railways, the nation's main transportation network, is currently drawing attention. This is mainly because of recent terrorist attacks aimed at private multipurpose facilities in a number of foreign countries. This article proposes a system for real-time monitoring of railway facilities and secure areas. Access control will be obtained using Raspberry Pi, an inexpensive micro-controller connected to the cloud via Amazon Web Service. Real-time surveillance is demonstrated by implementing computer vision and deep learning, and Twilio API. Intruders in restricted areas (such as tracks and electrical installations) can be detected with high precision and notifications can be sent to the safety and security managers in real time via short message service through cloud applications. The proposed system will assist the safety and security managers in responding swiftly and effectively to prevent or minimize risks that arise due to intruders.