Optimizing HVAC Efficiency via Deep Neural Networks for Real-Time Classroom Occupancy

Accurately determining the number of occupants in a room is crucial for optimizing smart environments and energy efficiency in HVAC systems. This paper presents a deep learning approach for precise, real-time classroom occupancy estimation to facilitate smart HVAC control. Utilizing a YOLOv4 object detection model, trained on an extensive dataset of labeled human faces, we developed a robust computer vison model with OpenCV libraries This model performs facial recognition and occupant counting through live video feeds from a Logitech c20 camera, achieving over 98% accuracy in typical classroom settings. We investigate the different techniques to address challenges such as occlusion and variability. The integration of our occupancy estimation model with HVAC control systems underscores a significant stride towards achieving energy conservation and sustainability goals in educational institutions, aligning with the emerging paradigms of smart building management systems.