Visibility inference based on spatial knowledge representation from observer's perspective

In this article, we propose a picture-based object-oriented spatial knowledge representation, called PCOS-string, for an environment seen from the observer's point of view. The PCOS-string representation preserves the spatial relationships among objects and thus can facilitate spatial reasoning especially from the observer's perspective. We also present a visibility inference algorithm based on the PCOS-string representation so that visibility of the objects in a given environment can be determined according to their spatial relationships. This visibility inference algorithm can handle both static and dynamic environments. Finally, we discuss the possibility of applying the PCOS-string and the associated visibility inference method to the mobile robot's path planning problem and virtual reality applications. © 1997 John Wiley & Sons, Inc.