Object Oriented Environment Model for Video Surveillance Systems

Conference paper


Thomas Emter
Ioana Gheta
Jürgen Beyerer


Klaus Thoma (ed.), Future Security: 3rd Security Research Conference, Fraunhofer IRB Verlag, 2008.




Future Security: 3rd Security Research Conference, Karlsruhe, September 10 - 11, 2008

The present contribution proposes a method to represent the relevant information extracted from a large number of sensors of a surveillance system, fused into a single comprehensive, dynamic model of the monitored area. The model is part of a semi-autonomous surveillance system N.E.S.T. developed at Fraunhofer Institute for Information and Data Processing - IITB. For building up the model, an object oriented approach based on Bayesian fusion is proposed, providing the system with a permanent situational awareness. The model also acts as an information hub for services responsible for inserting, evaluating and updating the information. The information in the model is condensed to represent only the relevant facts by means of fusion mechanisms, thus providing support for the surveillance team.