摘要：Recently, BEMS(Building Energy Management System) are emerging in the field of energy savings. Energy management approaches for Unmanned Building often become unreliable in complex building controls due to False detection, lighting changes, and other factors. We present a new framework to robustly and efficiently to detect people staying and going out based on background subtraction and foreground analysis with complement of tracking to reduce false positives. In our proposed system, the foreground is modeled by Adaptive Gaussian mixtures. In order to handle complex situations, several improvements are implemented for quick lighting change adaptation, shadow removal, fragment reduction, and keeping a stable update rate for video streams with different frame rates. Furthermore, the types of the people staying and going out are determined by using a method that exploits context information about the foreground masks, which significantly outperforms previous edge-based techniques. Lights dimming controls is also integrated to control multiple lights via a single system. The test and evaluation demonstrate our method is efficient to run in real-time while being robust to quick lighting changes and occlusions in real office environments.
