Geospatial data capture and handling of indoor spaces is increasing over the years and has had a varied history of data sources ranging from architectural and building drawings to indoor data acquisition approaches. While these have been more data format and information driven primarily for the physical representation of spaces, it is important to note that many applications look for the semantic information to be made available. This paper proposes a space classification model leading to an ontology for indoor spaces that accounts for both the semantic and geometric characteristics of the spaces. Further, a Space semantic model is defined, based on this ontology, which can then be used appropriately in multiple applications. To demonstrate the utility of the model, we also present an extension to the IndoorGML data standard with a set of proposed classes that can help capture both the syntactic and semantic components of the model. It is expected that these proposed classes can be appropriately harnessed for use in diverse applications ranging from indoor data visualization to more user customised building evacuation path planning with a semantic overtone.

Recent years has seen an increase in the work done on indoor data mapping and modeling. The standard data models provide different ways to store and access the indoor data but the way it is done is specific to the domain in which they are used. Although models like IFC, CityGML and IndoorGML provides rich functionality, the widespread availability of indoor data is not in these formats. This paper presents a step by step methodology to convert indoor building data of existing buildings, represented in architectural drawings into a topologically consistent and semantically rich indoor spatial model. The workflow presented consists of extracting relevant geometric entities from CAD drawings, assessing their topological relationships, using it to derive semantic information of spaces and making the data available in the form of IndoorGML. Since the current IndoorGML features lack the capability to store relevant semantic information, a semantic extension to IndoorGML is also proposed. The extraction of primitive spatial elements in rectilinear buildings like walls and doors are considered for the work presented in this paper. Development of a toolkit which implements this methodology in a seamless manner is work in progress and would incorporate extraction of complex spatial elements like staircases, ramps, curvilinear walls and windows, which is out of scope of the current work presented in this paper.

There have been various ways in which the indoor space of a building has been defined. In most of the cases the models have specific purpose on which they focus such as facility management, visualisation or navigation. The focus of our work is to define semantics of a model which can incorporate different aspects of the space within a building without losing any information provided by the data model. In this paper we have suggested a model which defines indoor space in terms of semantic and syntactic features. Each feature belongs to a particular class and based on the class, has a set of properties associated with it. The purpose is to capture properties like geometry, topology and semantic information like name, function and capacity of the space from a real world data model. The features which define the space are determined using the geometric information and the classes are assigned based on the relationships like connectivity, openings and function of the space. The ontology of the classes of the feature set defined will be discussed in the paper.