• Version
• Download 12
• File Size 316.98 KB
• File Count 1
• Create Date 17 March 2026
• Last Updated 17 March 2026

Real-Time Emergency and Alert Response System

HARIPRASATH . S
Departement Of Computer Science-PG
Kongunadu Arts and Science College
Coimbatore, india

Abstract:The rise in violent incidents in public spaces like workplaces, schools, transit hubs, and urban areas has made public safety a serious concern. Conventional surveillance systems depend on constant human observation, which is ineffective and vulnerable to human mistake, weariness, and delayed replies. In order to automatically identify violent activity from video streams, this study suggests a Real-Time Emergency and Alert System that uses deep learning and computer vision techniques. The system analyzes video frames and finds patterns associated to violence by integrating the YOLOv8 object detection technique with Python and OpenCV. By ensuring that alarms are only activated when violence is consistently recognized across consecutive frames, a multi-frame validation system lowers the number of false positives. The system creates notifications for security staff and takes screenshots of evidence as soon as an emergency is verified. Byconverting conventional passive monitoring into an intelligent automated security solution, the suggested method increases surveillance efficiency. Applications in smart campuses, public transportation systems, and smart city infrastructures can benefit from the architecture's support for both offline video analysis and live CCTV surveillance.
Keywords:Deep learning, computer vision, YOLOv8, real-time surveillance, and emergency alarm systems.