A robust occupancy detection and tracking algorithm for the automatic monitoring and commissioning of a building

Abstract

The focus of this study is the 24 h a day monitoring of buildings for commissioning purposes. Based on an image-based depth sensor and a programmable pan-tilt-zoom (PTZ) camera, the proposed monitoring system enables the continuous detection and tracking of the occupants, even under dim-lighting conditions. The proposed SVM-based observation measurement provides a more reliable tracking performance. This paper presents a robust day-and-night people tracking and counting algorithm. The function of large-scale field monitoring is realized using a PTZ camera network instead of a conventional fixed camera. Furthermore, based on the depth image sensor, the contour information of the occupant can be applied for more accurate activity recognition. In our experiments we demonstrated the positive result of the occupancy detection and tracking algorithms applied to count people and monitor a building.