Smart Surveillance Systems Using YOLOv8: A Scalable Approach for Crowd and Threat Detection

Abstract

In an era marked by increasing urbanization, public gatherings, and security risks, there is a growing demand for intelligent surveillance systems capable of proactive monitoring and threat mitigation. Traditional CCTV systems, which rely heavily on manual supervision, are limited by human fatigue, delayed responses, and inefficiency in detecting dynamic threats. This research introduces a comprehensive AI and ML-based surveillance framework designed to enhance public safety, crime prevention, and workplace monitoring using existing CCTV infrastructure. Leveraging the state-of-the-art YOLOv8 object detection model, the system enables accurate real-time detection, tracking, and classification of individuals and activities in live video feeds. It automatically monitors crowd density, detects suspicious behavior through behavioral analysis and anomaly detection algorithms, and generates timely alerts to aid rapid intervention by security personnel. The integration of deep learning techniques, such as convolutional neural networks and LSTM-based sequence models, ensures precise identification of deviations from normal behavior in both public and restricted zones. A significant emphasis is placed on minimizing false positives and computational overhead, making the system suitable for deployment on low-power edge devices.The proposed solution is further equipped with smart analytics, visual dashboards, and performance evaluation modules that assess model accuracy, precision, recall, and real-time responsiveness. Experimental results show that the system achieves 95.4% accuracy in object detection and 92.7% accuracy in anomaly recognition while reducing false alerts through context-aware filtering. Use cases span crowd control in public venues, industrial compliance tracking, traffic surveillance, and law enforcement applications. By offering a scalable, cost-efficient, and autonomous monitoring solution, this AI-powered system represents a transformative step toward smart surveillance and intelligent urban infrastructure.