Abstract: In weapon detection the Security remains a paramount concern across various domains, especially with the escalating crime rates in densely populated events or secluded areas. Leveraging computer vision for abnormal detection and monitoring presents significant applications in addressing numerous challenges. Given the increasing demand for safeguarding safety, security, and personal assets, the implementation and deployment of video surveillance systems capable of recognizing and interpreting scenes and anomalous events play a pivotal role in intelligence monitoring. This project proposes an automatic gun detection system using a YOLO (You Only Look Once) convolutional neural network (CNN)-based algorithm. The trained model exhibits the capability to detect guns based on a pre-trained YOLO file, triggering alerts via a buzzer and notifying preset authorized users or police stations. In addition, in the event of a threat, the victim can activate an emergency alert by pressing a button and vocally requesting assistance. This voice prompt is recognized, prompting an immediate alert with the captured scene for swift response.

| DOI: 10.17148/IARJSET.2024.11490