Chengzhi Rao

  • Current role:
    Geospatial Data Analyst at the "Global Dynamic Exposure Model" Group, Section 2.6 Seismic Hazard and Risk Dynamics, GFZ Helmholtz Centre for Geosciences, Germany

  • Research Interests:
    1. Geospatial data processing (Global building data, 3D model, 3D point clouds, Digital terrain model, Remote sensing, Transportation, Geospatial open data, etc)
    2. 3D Modelling for the built environment by using geospatial data: 3D building modelling, 2D and 3D road network modelling, etc
    3. Urban analytic

  • Education:
    1. Master of Geomatics, Delft University of Technology, the Netherlands (2022-2024)
    2. Master of Landscape Architecture and Urban Planning, Beijing Forestry University, China (2015-2018)
    3. Bachelor of Landscape Architecture, Beijing Forestry University, China (2011-2015)


Session

08-29
11:50
35min
Reconstructing A High-detailed Lane-Level Road Network Model from OpenStreetMap: A Connectivity-Driven Approach
Chengzhi Rao

OpenStreetMap centerlines carry enough topological, geometric and semantic information to reconstruct lane-level road networks — if interpreted correctly. This study proposes treating OSM centerlines as structural representations of dominant traffic flow rather than geometric midlines, and reconstructs detailed lane networks and 2D areal road geometries through a hierarchical, connectivity-driven process guided by topology, geometry, and OSM tag semantics. Applied to Delft, the Netherlands, the framework produces a complete, functionally typed lane-level network and seamless areal road representations covering diverse urban road configurations, validated against satellite imagery.

Martinique