Best Paper Award auf der ISPRS Geospatial Week
Das Paper DATABASE-SUPPORTED CHANGE ANALYSIS AND QUALITY EVALUATION OF OPENSTREETMAP DATA (Autoren: Alexander Martini, Paul V. Kuper, Martin Breunig) ist auf der internationalen ISPRS Geospatial Week in Enschede, NL mit dem Best Paper Award ausgezeichnet worden.
Die Veröffentlichung ist auf Basis der Masterarbeit von GuG-Student Alexander Martini entstanden, der seine Abschlussarbeit in Zusammenarbeit mit der Firma Disy Informationssysteme geschrieben hat. Betreut wurde die Arbeit an dem Lehrstuhl für Geoinformatik von Prof. Dr. Martin Breunig und Dr.-Ing. Paul Kuper. Auf der ISPRS Geospatial Week vorgetragen hat Paul Kuper.
Inhaltlich geht es in dem Artikel um die raum-zeitliche Datenanalyse und die Klassifikation von Änderungen für OpenStreetMap Daten.
Abstract:
A significant advantage of OpenStreetMap data is its up-to-dateness. However, for rural and city planning, it is also of importance to access historical data and to compare the changes between new and old versions of the same area. This paper first introduces into a differentiated classification of changes on OpenStreetMap data sets. Then a methodology for an automated database-supported analysis of changes is presented. Beyond the information already provided from the OpenStreetMap server, we present a more detailed analysis with derived information. Based on this approach it is possible to identify objects with attributive or geometric changes and to find out how they exactly differ from their previous versions. The analysis shows in which regions mappers were active during a certain time interval. Furthermore, a time based approach based on various parameters to determine the quality of the data is presented. It provides a guideline of data quality and works without any reference data. Therefore, an indication about the development of OpenStreetMap in terms of completeness and correctness of the data in different regions is given. Finally, a conclusion and an outlook on open research questions are presented.
 |