Urban models extracted from airborne data have to be refined for tasks like the generation of realistic visualisations from pedestrian viewpoints. Within the paper, terrestrial LIDAR data as well as façade imagery is used to increase the quality and amount of detail for the respective 3D building models. These models as they are available from airborne data collection provide a priori information, which can be integrated efficiently both for the georeferencing of the terrestrial data and the subsequent geometric refinement. After alignment of the terrestrial data to the given 3D model, window structures are first extracted approximately from the LIDAR point clouds. These structures are then further refined by 3D edges which are extracted from the overlapping façade images. Our modelling process applies a 3D object representation by cell decomposition, which can be used efficiently for building reconstruction at different scales.