A New Target Detection Scheme for Classroom Energy-saving System Based on YOLOv5s

Aiming at the problems of using infrared pyroelectric sensors to identify and detect human bodies in traditional classroom energy-saving systems, there are problems such as susceptibility to environmental temperature, difficult wiring, high financial resources, and high maintenance costs. A YOLOv5s-based classroom energy-saving target detection is designed. Plan. This solution can use the existing camera in the classroom without changing the original classroom wiring, and use the deep learning method to realize the detection of human body and lamps. At the same time, data enhancement, K-means generating a priori boxes, NMS (Non-Maximum Suppression) and other mechanisms are introduced to the network model, and the target can be detected while counting the number of targets and outputting the positioning information of the target. Compared with the traditional single human body detection, the difference is that this solution also realizes the detection of the lighting status of the lamps. The experimental results show that the proposed model is applied to the classroom energy-saving system target detection with high detection accuracy and high recall rate, that is, the model has strong robustness and generalization ability, and can be effectively applied to the classroom energy-saving system.