Multi-hierarchical Quality Assessment of Geospatial Data

In this paper we describe the analysis framework system GEOAIDA with view of an application for a versatile and efficient quality assessment of geodata. GEOAIDA allows to develop image analysis strategies for complex object class definitions being provided by different GIS databases. The image analysis strategy and GIS data model can be expressed in tree-like semantic networks. The multi-hierarchical architecture allows multiple combinations of image analysis tools for a multifaceted use. A practical application is the semiautomatic quality assessment of MGCP data by using IKONOS imagery. For the comparision of MGCP and image content, information about different object classes is extracted from the image. The landcover objects are detected by color texture classification combined with structural analyzing methods. Roads are detected by a line based extraction algorithm combined with color texture classification results. The MGCP database itself is used as prior knowledge to perform the image analysis.