Enhancing Security with Automated Weapon Detection: A CNN and YOLO-V5-Based Video Surveillance System with Alarm Integration

The rising number of gun-related incidents worldwide has brought the need for a better and trustable weapon detection and alarm system to improve security. Although several weapon detection systems based on Machine learning, AI and many other techniques have been proposed, they often suffer from high error probability and slow real-time detection, leading to a high rate of false negatives. This research proposes a video surveillance system based on You Only Look Once (YOLO) to address these challenges. The proposed system aims to minimize false positives and false negatives while maintaining fast detection speed. This system also has an alarm integration that alerts the user when the weapon is detected. The results indicate that the proposed YOLO-v5-based system achieves a low error rate and fast detection speed, making it a better solution for real-time weapon detection in security applications. The system is better suited for the security surveillance as it provides better security. Overall, this system makes a better surveillance system in terms of weapon detection