Computer Vision-Based Intelligent Classroom Systems for Efficient Power Management in Large Educational Institutions

Abstract

The increasing demand for electricity coupled with depleting resources necessitates urgent action to reduce wastage and address the pressing need for efficient energy utilization. In large educational institutions a huge amount of energy is being wasted as the electrical appliances of a classroom are often switched on even though there are no occupants in the room. This thesis explores the development of an intelligent classroom system to enhance energy efficiency and reduce electric bills in traditional classroom settings. By leveraging automation and Internet of Things (IoT) technologies, the research aims to optimize energy consumption by tracking real time occupancy and controlling lighting and fans accordingly. Various occupant detection models, including the MOG2 algorithm, OpenCV, and YOLOv3, are compared to identify the most effective approach for accurately detecting human occupants. The study presents a comprehensive methodology that includes the design of a hardware setup using Raspberry Pi, Pi cameras, and other components for occupant detection and appliance control. The key finding underscores the superior performance of YOLOv3 in accurately identifying human occupants while minimizing false detections. The proposed intelligent classroom system offers a user-friendly and customizable solution, enabling efficient energy management and cost reduction. This research contributes to the advancement of smart classroom technologies, promoting sustainability and improved resource utilization in educational institutions.