Towards generating highly detailed 3D CityGML models from OpenStreetMap

Abstract

About one decade has passed since US vice president Al Gore articulated his vision of Digital Earth (DE). Within this decade, a global multi-resolution and three-dimensional (3D) representation of the Earth, which sums up the DE vision, increasingly gained interest in both public and science. Due to the desired high resolution of the available data, highly detailed 3D city models comprise a huge part of DE and they are becoming an essential and useful tool for a range of different applications. In the past as well as at present, 3D models normally come from a range of different sources generated by professionals, such as laser scans or photogrammetry combined with 2D cadaster data. Some models are generated with semi-automated or fully automated approaches, but in most cases manual fine tuning or even manual construction from architectural plans is required. Further beyond outdoor city models, DE additionally envisages the provision of indoor information. That is, the interior structure of public or publically accessible buildings, such as airports or shopping malls, is represented and made available in 3D; however, at the moment, such models are mostly created by hand and essentially based on professional data sources. In contrast to such professional data, which is mainly captured by surveyors or companies, the last few years revealed the phenomenon of crowdsourced geodata, which receives an increasing attractiveness as an alternative data source for many Geographic Information Systems (GIS). Former research already demonstrated the power and richness of such geodataespecially OpenStreetMap (OSM) -and it has also been proved that this non-standardized, crowdsourced geodata can be combined with international standards of the Open Geospatial Consortium (OGC). For example, CityGML Level-of-Detail 1 (LoD1) and LoD2 models have already been created automatically from OSM. The research presented in this article will further continue on the automated generation of CityGML models from OpenStreetMap. Essentially, a method for the creation of highly detailed CityGML LoD4 models with interior structures will be explained. By applying the invented approach on existing OSM data, limitations and restrictions of the IndoorOSM mapping proposal, the available data and the developed approach are revealed and discussed.

Formalisation of the level of detail in 3D city modelling

“…Forberg (2007) expresses that it as a common way to enhance the performance of interactive visualisation of polyhedral data. According to Sester (2007) and Goetz (2013) LODs are multi-scale models for different applications. Lemmens (2011) equals it to the term of resolution and states that it is related to how much detail is present in the data and may refer to space, time and semantics.…”

An improved LOD specification for 3D building models

“…They are normally derived from terrestrial laser scanning (El Meouche et al, 2013;Akmalia et al, 2014), very dense airborne laser scanning (Truong-Hong and Laefer, 2015), their combination (Kedzierski and Fryskowska, 2014), from the conversion and generalisation from architecturally detailed models such as BIM (Donkers et al, 2015;Geiger et al, 2015;Isikdag and Zlatanova, 2009;de Laat and van Berlo, 2011) and CAD (Lewis and Séquin, 1998;Huang et al, 2008), from architectural plans (Yin et al, 2009), ground imagery (Xiao et al, 2009), and with procedural modelling (Goetz, 2013;Smelik et al, 2014;Martinović et al, 2015). Recent research in the acquisition of LOD3 models is focused to automatisation, especially automatic detection of windows and other façade details (Becker, 2009(Becker, , 2011Van Gool and Martinović, 2013).…”

Processing BIM and GIS Models in Practice: Experiences and Recommendations from a GeoBIM Project in The Netherlands

“…LOD0 is non-volumetric and is an horizontal footprint and/or roof surface representation for buildings; LOD1 is a block-shaped model of a building (with an horizontal roof); LOD2 adds a generalised roof and installations such as balconies; LOD3 adds, among others, windows, doors, and a full architectural exterior; LOD4 models the interior of the building, potentially with pieces of furniture (CityGML does not mandate which indoor features need to be modelled, in practice resulting in models with a different granularity [10,11]) (LOD4 will be removed in CityGML 3.0. Instead, indoor and outdoor features will each be modelled at LOD0-3.).…”