Extending OpenStreetMap to indoor environments: Bringing volunteered geographic information to the next level

Extensive and high-quality geographic data sources are important for any kind of spatial analysis or application, especially in the field of urban data management. In the last couple of years, Volunteered Geographic Information (VGI) has increasingly gained attractiveness not only to amateur users, but also to professionals in the geoinformation industry. Different VGI communities evolved and especially the OpenStreetMap (OSM) community became very strong. OSM provides very detailed information about the landscape, the street network and also more and more buildings are mapped. However, until now, this building mapping is mainly related to the outer shape of the ground space of the building and there is hardly any information about the inner structure available. This paper presents an approach of extending the OSM tagging schema to indoor environments. A 3D Building Ontology targeted to VGI communities is presented for describing different information aspects about buildings and their inner structure. Based on this ontology, the OSM extension is developed and explained. A proof of concept is given by applying the developed extension in mapping a use case building. Diverse online mapping platforms such as OpenStreetMap (OSM) or Wikimapia have been initiated, allowing users to contribute and collaboratively edit spatial data. The increasing number of participants leads to a variety of different spatial data and information about geographic phenomena such as the street network, cities, POIs, buildings, landuse etc. At UDMS 2009, VGI has also been shown to be useful for urban data management by providing new sources of information that even can be integrated into a 3D platform (Schilling et. al. 2009). However, most of current mapping activities are related to the outdoor environments (e.g. rough building structures, information about type of businesses, playgrounds, footpaths etc.) and there is hardly any information about indoor environments available. Since there is an increasing need for mature indoor navigation solutions and other indoor location based services (Goetz & Zipf 2010) and data providers are hardly able to commercially capture indoor data for large areas, there is an enormous potential within VGI communities for capturing and providing information about indoor environments which are open to the public (e.g. airports or shopping malls). For that purpose, it is essential to provide clear and understandable methodologies for mapping data about indoor environments. Therefore the main contribution of this paper is an extension to indoor environments for bringing VGI to the next level. The remainder of this paper is organized as follows: First, there is an introduction to OpenStreetMap, especially focusing on its data model and data acquisition methods. Afterwards, there is a brief overview about related research. Thereafter, an extensive ontology for 3D building models with detailed information about indoor environments is presented. Subsequently, there is a description of how to extend the existing OSM tagging schema to indoor environments according to this ontology. The presented methodology is demonstrated on a use-case building and the last chapter summarizes the presented research and discusses future work. 2 THE OPENSTREETMAP COMMUNITY One very popular (or even the most popular) VGI community is the so called OpenStreetMap project. OSM follows the peer production model (Haklay & Weber 2008) that created Wikipedia and aims for the provision of free to use and editable map data. Since 2004, the project grew rapidly and by November 2010 there were more than 320.000 registered users and more than 2.000 millions tracked points in the database (OSM 2010a). The data in OSM is created in different ways. The most important way is the acquisition of original data, manually captured by users via GPS devices. However, people can also contribute data based on aerial images (e.g. by Bing or Yahoo) or by contributing their local knowledge about the region they live in. Basically, OSM consist of differently tagged nodes, i.e. a point with distinct coordinates. For defining lines or simple polygons (i.e. a polygon without holes), users can create so called ways which consist of several nodes. For defining a polygon, this way needs to be closed, i.e. the start point equals the end point. For mapping complex polygons or describing existing relationships between different elements, there are furthermore so called relations. These contain ways, nodes and also other relations. In conjunction with user-generated content and collaboratively collected data, there is always a question about the accuracy and quality of the provided data. By comparing OSM data with data provided by commercial vendors like Teleatlas, it became evident that VGI is able to compete against commercial providers (Cf. (Zielstra & Zipf 2010), (Haklay 2010), (Ludwig et. al. 2010)). That is, data from OSM can be considered as a real alternative data source for spatial and geographic data in urban environments. The data of OSM mainly focuses on outdoor environments and objects (e.g. streets, landuse etc.). When considering buildings, some information can be available within OSM, but these do only refer to very basic things like the location, outer shape or the height of the building. The latter mentioned information can be applied in the 3D visualization of city models as within the OSM3D project (Over et. al. 2010). There are also some discussions about indoor mapping (Cf. next chapter), but until now they cannot be regarded as mature.