Automated vigilence system for physical clash detection in localized environments

Using MobileNetV3, a compact framework for neural networks and computer vision techniques, this study offers a proposal on how to automate vigilance systems for physical clash detection for continuous monitoring in real time. The primary goal of this system’s design was to identify probable physical confrontations between nearby individuals and automatically make an audio message to the appropriate authorities to report a case like this. The system is trained using a real-time violence. Situations video dataset from the MobileNetV3 big model. Through a variety of video inputs, the model’s ability to recognize physical collisions is tested, and it demonstrated a precision of 96%. Real-time use of the model’s saved weights is made possible by a live video feed, and a phone call alarm is automatically sent to the registered authority using the Twilio API so that immediate action can be taken to prevent more physical encounters. Camera surveillance can be used to deploy the system. As a result, the suggested method represents an important advancement in the field of automated monitoring and can be seen as making a substantial contribution to enhancing security. Overall, this paper highlights the effectiveness of keeping tabs on locations for physical clash detection and the possibility of automated surveillance devices.