A Real Time Human Detection System Based on Far Infrared Vision  Yannick Benezeth 1 , Bruno Emile 1 , H´el`ene Laurent 1 , and Christophe Rosenberger 2 1 Institut Prisme, ENSI de Bourges - Universit´e d’Orl´eans - 88 boulevard Lahitolle, 18020 Bourges cedex - France 2 Laboratoire GREYC, ENSICAEN - Universit´e de Caen - CNRS, 6 boulevard Mar´echal Juin, 14000 Caen - France yannick.benezeth@ensi-bourges.fr Abstract. We present in this article a human detection and tracking algorithm using infrared vision in order to have reliable information on a room occupation. We intend to use this information to limit ener- getic consumption (light, heating). We perform first, a foreground seg- mentation with a Gaussian background model. A tracking step based on connected components intersections allows to collect information on 2D displacements of moving objects in the image plane. A classifica- tion based on a cascade of boosted classifiers is used for the recognition. Experimental results show the efficiency of the proposed algorithm. 1 Introduction Vision based systems can nowadays be found in many applications, for monitor- ing goods in private areas or for managing security in public ones. Nevertheless, the relative robustness of vision algorithms, the camera miniaturization and the computation capacity of embedded systems permit other applications of vision based sensors. In order to keep at home low mobility people or to manage the energetic consumption, we need some reliable information on room occupation, the number and the activities of the house occupants. Vision based systems are probably the most efficient technology for this task. The Capthom project, in which we are working, falls within this context. It consists in developing a human detection low cost sensor. This sensor will present advantages compared to existing sensors, that is to say, a strong immunity to intemperate detections and a great detection reliability. We want to have a refer- ence platform which can give information on a room occupation. This platform will assess the performance of other affordable technologies (e.g. cameras in the visible spectrum). In spite of its prohibitive price, far infrared technology is the most convenient one to automatically watch a room. Acquisition is not influ- enced by the luminosity. In this framework, we have developed a far infrared algorithm which can detect and track a human in a room.  This work was realized with the financial help of the Regional Council of Le Centre and the French Industry Ministry within the Capthom project of the Competitive- ness Pole S 2 E 2 . A. Elmoataz et al. (Eds.): ICISP 2008, LNCS 5099, pp. 76–84, 2008. c  Springer-Verlag Berlin Heidelberg 2008