AUTONOMOUS FIREARMS DETECTION AND THREAT MONITORING USING DEEP VISION MODELS IN SMART CITIES

Abstract

Ensuring public safety in rapidly urbanizing environments requires advanced surveillance systems capable of detecting threats in real time. Firearms-related violence poses a serious risk to smart cities, where manual monitoring is often inefficient and prone to human error. This research proposes an autonomous firearms detection and threat monitoring framework based on deep vision models. The system employs state-of-the-art deep learning algorithms such as YOLOv8 and EfficientNet to identify and classify firearms from live video streams with high precision. An adaptive alert module automatically notifies law enforcement authorities upon detection, minimizing response time. The model is trained on a diverse dataset comprising various firearm types and environmental conditions to ensure robustness across scenarios. Experimental results demonstrate superior detection accuracy, scalability, and low false alarm rates compared to traditional surveillance techniques. This study highlights the transformative potential of deep vision models in creating safer, intelligent urban infrastructures.