An Automated CityGML LoD1 3D City Modeling with UAV Technology and FME Scripting

The increasing demand from industry and governance adds to the importance of introducing new ways of 3D City Modeling. The current study presents an advance automated 3D City Modeling technique using the concept of Unmanned Aerial Technology and FME Scripting. A detailed study presents step by step procedure of achieving a CityGML standard LoD1 3D City Model. First part of the current study implements a commercial grade UAV equipped with optical camera for overlapped image acquisition in the selected urban slum area of Roorkee, India. Overlapped images are processed with UAV data processing software, Pix4D for orthoimage generation and digital surface modelling. Very high resolution orthoimage is segmented using state of art Quick Shift Segmentation scheme. The segmented image further classified using Support Vector Machine classifier for building footprint extraction. In the second part of the study, FME 'LoD1 3D City Model Extraction Framework' is proposed to extrude building footprint (ESRI Shape file (.shp file) format) to the corresponding height values (separate excel sheet) computed using digital surface model and ESRI zonal statistics toolkit. This script helps to overcome a long felt gap in automated 3D City Modeling practices and strengthen the pathway for developing 3D GIS spatial analytical techniques and tools.