Image Processing Based Controller for Indoor Cooling Unit

Abstract

For interior cooling systems, efficient energy management is crucial to reducing power consumption and running expenses. In this work, an image-processing-based controller that uses real-time occupancy identification and tracking to optimize air conditioning utilization is presented. To precisely count the number of people in an area, the system analyzes CCTV camera footage using YOLOv5 for object detection and Deep SORT for multi object tracking. To increase efficiency, the monitored area is separated into many areas, each of which is managed by a separate air conditioner. The system allows for automated control using the MQTT protocol, guaranteeing that cooling units only run when a designated area is occupied and dynamically modifying the temperature in response to occupancy density. According to experimental data, the suggested method greatly reduces energy loss while achieving excellent occupancy detection accuracy. By providing a scalable and economical smart building management solution, this approach promotes sustainable operations and energy reduction.