YOLOv5-Based AI-Enabled IoT and Cloud Computing with SDN for Real Time Weapon Detection in Video Surveillance

Abstract

Weapon detection in real-time surveillance systems is crucial for public safety and crime prevention. This paper 
proposes an AI-enabled IoT and Cloud Computing framework integrated with Software-Defined Networking 
using YOLOv5 for real-time weapon detection in video surveillance. The proposed system employs deep 
learning-based object detection with optimized network traffic handling to enhance computational efficiency 
and response time. Extensive experimentation was conducted using the SOHAS Weapon Detection dataset, 
achieving an impressive mean Average Precision of 97.3%, precision of 96.8%, recall of 95.6%, and an F1
score of 96.2%. Comparative analysis demonstrates superior performance over existing methodologies in terms 
of detection accuracy and real-time processing efficiency. The framework's integration with SDN improves 
network adaptability, reducing latency by 28% and increasing throughput by 35%. Furthermore, IoT-enabled 
edge devices ensure seamless data transmission, enhancing surveillance effectiveness. This hybrid approach 
overcomes limitations in traditional surveillance systems, offering a scalable, high-performance solution for 
real-world applications. The results indicate the robustness of the proposed system in high-traffic surveillance 
environments, ensuring reliable weapon detection with minimal false positives. Future work will focus on 
expanding the dataset and optimizing computational resources for large-scale deployment.