Artificial Intelligence in Cyber-Physical Security and Intelligent Surveillance: A Comprehensive Survey

Abstract

Modern times require smart solutions for crime detection to enhance public security levels because crime detection remains a fundamental challenge. Traditional surveillance relies on human monitoring but has weaknesses due to time consumption and human error. The automated crime detection system runs on machine learning and deep learning platforms using Convolutional Neural Networks and Recurrent Neural Networks, along with YOLO, Faster R-CNN, and the Azure Face API. The AI models perform two functions: detecting suspicious activities and identifying weapons and faces, and recognising patterns of criminal behaviour. The combination of local servers and cloud platforms powered by AI detects crimes precisely in real time through enhanced monitoring, improves accuracy, and prevents false alarms, while also enabling broader crime-prevention capabilities.