Change Detection in CityGML Documents

With the increasing importance and usage of virtual 3D city models, a suitable continuation process becomes necessary to ensure the sustainability of 3D city models. If a new model was gathered e.g. by processing from LIDAR data or new data acquisition—knowledge about the differences between the old and the new city model is beneficial. This would allow the versioning, update and historicization. In this contribution we will present a concept for the detection of changes between two CityGML instance documents to ensure sustainable continuation processes in the field of 3D city models. The proposed concept will be based on techniques of existing change detection algorithms and object detection algorithms from the field of image analysis. The detected changes will be saved in a well-defined format so that following applications such as database operations can be performed or can make use of them. Finally, we examine the performance of our proof-of-concept implementation.