Vision-based indoor occupants detection system for intelligent buildings

Abstract

In intelligent buildings, practical sensing systems designed to gather indoor occupancy information play an indispensable role in improving occupant comfort and energy efficiency by optimizing control strategies of HVAC (Heating, Ventilation and Air Conditioning) system and lighting system. In this paper we propose a novel method for occupant detection based on video surveillances now widely used in buildings. In our method, a two-staged static detector using both Haar-like and HOG (Histograms of Oriented Gradients) features and a template-based motion analysis module are concatenated to detect the heads of occupants rapidly and effectively. The accuracy can satisfy the requirements of the automation systems in intelligent buildings. Experimental results demonstrate the effectiveness and efficiency of the proposed method.