Abstract

In this paper a knowledge-based approach to the interpretation of public utility maps is presented. The interpretation is based on two types of a priori knowledge which are represented in a semantic network. The semantic network provides an efficient representation for reasoning with a priori knowledge and interpretation of results. An interpretation system for maps based on the semantic network is described. In this system a top-down control mechanism is integrated with a bottom-up object recognition strategy. The interpretation system has been tested on public utility maps, and the results are presented and evaluated.