Semi-automatic quality control of topographic data sets

The usefulness and acceptance of geo-information systems are mainly depends on the quality of the underlying geo-data. This paper describes a novel system for semiautomatic quality control of existing topographic geo-spatial data via automatic image analysis. The goal is to reduce the manual effort for quality control of a GIS database to a minimum. The core of the system is a semantic network in which different image analysis operators can be included. The image analysis operators are created for specific applications, i.e., the quality control of specific object classes which are most relevant. Images which can be used in the system are aerial images, high-resolution satellite imagery, and low-resolution satellite imagery. A prototype of the system has been in use for several years at public mapping organizations. From the experience gained during this time, we give a detailed report on the system performance and an evaluation of the results.