Change detection of buildings


With the temporal ADE for CityGML (presented here) that proposes time management for 3D urban data contained in a CityGML file, we are able to provide dynamic visualization of the city and its evolution over time by using a file containing city objects constituted of states defined by different dates. However, this requires us to prepare the data since such models containing different states each linked to date are not directly accessible. We initially have a set of distinct and independent CityGML files representing the same urban area, but generated at different dates. We must group them in a model containing all 3D objects defined in the multiple files and each of them should contain temporal information describing its evolution over the period represented by all these files. We worked on the realization of this model for buildings by comparing pairs of input models.

Considering that we have two files representing the 3D buildings of a city according to two different acquisition dates, we have developed a method to detect changes that occurred between them. By comparing the buildings of these two models, we are able to classify them into four categories: those which have been destroyed, constructed, modified, or remained unchanged between the two dates. Thanks to this classification, we can generate a CityGML model containing various buildings and temporal information corresponding to their category.


Videos


Publications

This method has been presented at the international conference “3D GeoInfo 2014”.

Pedrinis F, Morel M, Gesquière G (2015) Change Detection of Cities. In 3D Geoinformation Science, Springer International Publishing, pp: 123-139