Threat Detection System

The increasing incidence of robbery, physical violence, and weapon misuse in public and private environments highlights the limitations of conventional CCTV-based surveillance systems that rely heavily on human monitoring. This paper presents the design and implementation of an AI-powered IoT-based intelligent surveillance system capable of real-time threat detection. The proposed system integrates low- cost ESP32-CAM and NodeMCU modules for image acquisition with a lightweight Convolutional Neural Network (CNN) optimized using TensorFlow Lite for efficient edge inference. The system detects suspicious activities such as robbery, physical fighting, and weapon presence, and sends real-time alerts through an Android application using Firebase services. The proposed solution is cost-effective, scalable, and suitable for deployment in banks, educational institutions, offices, and public areas.