Investigating result quality indicators in industry foundation classes to city geography markup language conversion

Abstract: Industry Foundation Classes (IFC) models from the Architecture, Engineering and Construction (AEC) -industries can contain information on building life-cycle entities. Urban planning and related domains are interested in data conversions between IFC and City Geography Markup Language (CityGML) because of this. However, the maintenance and versioning of CityGML life-cycle building data poses a serious challenge. Semantics play a big role in CityGML data contents in addition to converted objects. General methods for testing data conversion result for quality are not strongly present in the prevailing research literature. The evaluation of CityGML file contents and conversion quality are therefore challenging. CityGML is a Three-dimensional (3D) data format created for storing 3D city data into databases. The exchange of life-cycle data from BIM environments enables more spatial analyses on urban and environmental related data. This master thesis explores methods that are in use for evaluating conversions and data accuracies within the realms of Building Information Modelling (BIM) and geodata by researching application requirements and measures for quality in a case study. The INSPIRE directive implementation guides distribution of public domain geodata in the EU and gives instructions on implementing CityGML. Different applications of CityGML and their requirements are leading to the creation of national guidelines. A literature study/review conducted revealed four different main sources for CityGML data requirements. An explorative case study compares four different IFC models from the GeoBIM benchmark testbench data. The conversion tools FME 2017 and ArcGIS PRO Data Interoperability extension are used to convert the GeoBIM benchmark IFC data to CityGML 2.0 Level-of-Detail (LOD) 3 and LOD4 data. A total of ten test methods are performed to assess the quality of IFC to CityGML conversion data. The evaluation results for the quality indicators created in the case study reveal that most metrics used for indicating quality of IFC to CityGML 2.0 data conversions are applicable for single LOD4 features but are more difficult to interpret for LOD3. The results from the study reveal that the conversion methodologies for IFC data should be verified before tackling performance and optimisation issues. Metrics for deriving positional accuracies within the data conversion geometries and those taking advantage of the FME data inspector features are easier to apply. Detailed findings from the case study data conversions revealed more interesting facts about the data evaluation methods and conversion workflows. There is a severe lack of automated eXtensible Markup Language (XML) formatters for writing CityGML. More study is also required on the documentation of data conversion methods.