Object-based Verification and Update of a Large-scale Topographic Database

It is well known that there is a growing need for consistent and up-to-date GIS-data at various scales. This paper describes a system for semiautomatic quality assessment and update of an existing large scale topographic database. The necessary reference information is derived from current digital aerial orthophotos with a ground sampling distance of 0.1m via automatic image analysis. The advantage of the system is that it reduces the manual efforts of a human operator to a minimum. To overcome the limitations of the automatic components the human interaction is focused on those objects, for which no reliable assessment result can be achieved within the automatic verification process. The efficiency of the whole system depends on the performance of the used image analysis component, because the number of correct objects for which not enough evidence can be found in the imagery is the limiting efficiency factor. The focus of the paper is on the assessment of the most important object classes of a large scale topographic database maintained at the Ministry of Municipal and Rural Affairs (MOMRA), Kingdom of Saudi Arabia. Two different applications are implemented and evaluated. The first is the verification of the existing GIS-data and the second is the update of the GIS-data. Both applications were implemented using the software GeoAIDA that is a so-called knowledge-based system which was developed at the Institut für Informationsverarbeitung (TNT), Leibniz Universität Hannover, Germany. The automatic image analysis is based on a supervised texture classification. Manual evaluation results of the verification and the update approach of the MOMRA GIS-data in different test regions are shown and the results are discussed.