[2H1-OS-28] OS-28

9:00 AM - 9:15 AM JST(12:00 AM - 12:15 AM UTC)

[2H1-OS-28-01] Proposal for a Hierarchical Road Auto-Generation Ecosystem for Wide-Area Maintenance and Continuous Data Updates

〇Hirofumi Omichi¹ (1. Silicon Studio Corporation)

Keywords:

Geospatial Information,Open Data,Road Auto-Generation,Wide Area,Sustainable Data Updates

3D city models, including Project PLATEAU, are increasingly being utilized across various fields such as autonomous driving simulations, urban development, and entertainment. However, LOD1 remains the mainstream for road models, while LOD2 which includes detailed road information like roadway and sidewalk sections is only available for limited areas. Furthermore, the high costs of both data creation and maintenance pose a significant burden, presenting challenges for sustained efforts. Therefore, this study proposes a system for automatically generating pseudo-LOD1 and LOD2 road models covering all of Japan, utilizing open data from the Geospatial Information Authority of Japan (GSI). This system is a QGIS-based, data-driven system aiming to create an ecosystem that combines open data for nationwide data deployment and sustainable data updates. Experimental results showed that the Intersection over Union (IoU) between the generated road model and the PLATEAU road model averaged 0.9739 for LOD1 and 0.8149 for LOD2.