AI-based Smart Surveillance for Behavioral Risk Face and Weapon Detection

Abstract

Modern surveillance systems play a crucial role in maintaining security in public places such as airports, railway stations, banks, and shopping malls. Traditional cctv surveillance mainly depends on continuous human monitoring, which is time-consuming and often inefficient when multiple cameras are involved. In many situations, security personnel may miss important events due to fatigue or human error, which can delay the response to potential threats. Therefore, there is an increasing need for intelligent surveillance systems that can automatically analyse video streams and detect suspicious activities in real time. This paper proposes an ai-based smart surveillance system for behavioural risk, face and weapon detection that uses computer vision and deep learning techniques to enhance public safety. The proposed system captures live video frames from a surveillance camera and processes them using advanced detection models. A face detection module is used to identify human faces in the video stream, and a face recognition mechanism can be applied to compare detected faces with stored records. At the same time, a weapon detection module based on the yolov8 object detection model analyses the frames to identify dangerous objects such as guns and knives. When a weapon is detected, the system automatically highlights the object using bounding boxes and generates alerts to notify security personnel. In this context, the proposed project introduces an ai-based smart surveillance system for behavioral risk, face and weapon detection. The system captures live video frames and processes them using computer vision techniques to detect human faces and identify weapons such as guns and knives. When a potential threat is detected, the system generates alerts and records the detection information for monitoring and investigation purposes