USING EXISTING CCTV NETWORK FOR CROWD MANAGEMENT, CRIME PREVENTION & WORK MONITORING USING AI & ML

Abstract

Traditional CCTV systems mainly record surveillance footage and depend on continuous human observation, limiting timely identification of crowd congestion, suspicious activities, and workplace safety issues. This paper presents an AI- and ML-enabled surveillance framework that enhances existing CCTV infrastructure without requiring major hardware replacement. The proposed system acquires live or recorded video streams, extracts frames, and performs preprocessing using OpenCV before applying YOLOv8 for real-time object detection. Detected objects are forwarded to crowd management, crime monitoring, and workplace monitoring modules. Crowd conditions are evaluated using configurable thresholds, while identified irregular or safety-related events are connected to an alert generation module and monitoring dashboard. Python, TensorFlow, NumPy, and OpenCV support the implementation and visualization workflow. The system is evaluated using unit, integration, functional, system, white-box, black-box, and acceptance testing. The contribution is a modular intelligent surveillance architecture that integrates object detection, operational monitoring, automated alerts, and reporting through existing CCTV infrastructure. Keywords:.Artificial Intelligence, Machine Learning, CCTV Surveillance, Computer Vision, YOLOv8, Crime Prevention, Workplace Monitoring, Real-Time Video Analytics.