Intelligent Public Surveillance System

The goal we propose in this paper is to develop an Intelligent Public Surveillance System to improve public safety and aiming for real time incident detection. Currently, traditional surveillance systems cannot handle the complexity of event detection and analysis over time in dynamic environments and thus may result in delayed or erroneous response. Our system utilizes advanced deep learning technology (Long-term Recurrent Convolutional Networks (LRCNs) for anomalous events detection, and YOLO (You Only Look Once) for real time population density estimation). They are the technologies that allow the detection of critical events such as fights, explosions and road accidents. We further introduce a user-friendly graphical user interface for administrators to monitor video streams, view analytics and manage incidents. In addition, the system uses MongoDB GridFS for scalable, efficient video storage and retrieval. The proposed system offers a solution to intelligent surveillance, real-time detection and actionable insights for public safety management.