Digital City Models · Geoinformatics · Automation
Radiate brings together projects for processing, modelling and automatically generating digital city models. The focus is on open standards, reproducible geodata workflows and specialised tools for CityGML.
Selected Work
Four development areas combine CityGML, Blender, remote sensing data and machine learning into integrated workflows for digital city models.
A Blender plugin for the georeferenced import, editing and export of CityGML building objects. The workflow connects standardised 3D city models with Blender's modelling and automation capabilities.
Add GitHub linkMachine-learning-based detection of windows and doors in building façades. The detections are geometrically refined, converted into openings and integrated into CityGML building objects.
Add GitHub linkAn automated workflow projects aerial images onto CityGML surfaces, evaluates suitable camera perspectives and creates textured city models with reproducible image assignment and texture atlas generation.
Add GitHub linkBuilding footprints, digital terrain models and laser scanning data are combined to automatically derive georeferenced CityGML building objects, including height determination, roof reconstruction and semantic surface generation.
Add GitHub linkFrom Source Data to City Model
The projects share one objective: to automate complex and recurring steps in 3D geodata processing in a transparent and reproducible way without losing sight of georeferencing, semantic structure or exchange formats.