YOLO-Driven Real-Time Centralized HVAC Temperature Monitoring via Autonomous Inspection Robots in Smart Factories

With the development of intelligent manufacturing, factory environment monitoring is gradually moving toward automation and intelligence. As a critical system for maintaining comfort and energy efficiency in production environments, the operational status and temperature control of central HVAC (Heating, Ventilation, and Air Conditioning) systems are critical. This paper proposes a real-time central HVAC temperature detection system based on the WCA-YOLO object detection algorithm, aiming to enhance temperature monitoring efficiency and coverage in intelligent production and office environments through autonomous inspection robots. Firstly, an HVAC temperature detection dataset was constructed, including 20 categories: 19 temperature levels corresponding to adjustable settings and a panel-off state. All images are meticulously annotated to ensure high-quality training data. Secondly, the WCA-YOLO model is built upon the YOLOv8 architecture, incorporating Wavelet Pooling to enhance sensitivity to edge information in images and adopting the Cross Stage Partial (CSP) mechanism to improve the detection performance for complex backgrounds and fine details. Finally, experimental results demonstrate that WCA-YOLO achieves superior performance in HVAC air conditioner detection tasks, establishing it as the optimal approach for this application.