REAL-TIME DETECTION OF PROHIBITED OBJECTS IN PUBLIC SPACES BASED ON IOT AND DEEP LEARNING

Donatien Kadima Muamba (Faculté des Sciences Informatiques, Université Révérend KIM)
David Muanza Lubukayi (Institut Supérieur Pédagogique)
Ali Akake Nengo (Faculté des Sciences Informatiques, Université Islamique au Congo)

IJISCS (International Journal of Information System and Computer Science)
Vol 10, No 1 (2026): IJISCS (International Journal of Information System and Computer Science)

Security in public spaces is a major challenge in the face of diversifying threats. Traditional video surveillance systems, relying on continuous human supervision, have limitations in terms of responsiveness and reliability. This article proposes an intelligent system for detecting prohibited objects, combining the Internet of Things (IoT) and deep learning. The architecture is based on an embedded ESP32-CAM module for image acquisition and a backend server using a deep learning model for analysis. Experimental results show an overall accuracy of 92.8%, demonstrating the suitability of this approach for real-time automated surveillance applications