Scientific Reports paper uses OpenStreetMap to power four urban planning models
anselm · x · 2026-07-28
A Scientific Reports paper describes a unified deep learning framework that combines OpenStreetMap with satellite imagery plus demographic and environmental data to solve four urban-planning tasks.
- Land-use classification: CNN, 91.6% accuracy
- Building footprint extraction: U-Net segmentation
- Traffic prediction: LSTM
- Air-quality forecasting: hybrid CNN-RNN
The key takeaway is that the shared map layer matters more than forcing one giant model to do everything. The authors argue OpenStreetMap has become a foundational geographic dataset that multiple AI systems can build on.
More from Research
- Chelsea Finn says robotics is bottlenecked by rollout cost and memory — chris_j_paxton · 2026-07-28
- This flow model learns both what to generate and how large it should be — KevinKaichuang · 2026-07-28
- Expanding Flow Maps can insert tokens or nodes while denoising — KevinKaichuang · 2026-07-28
- Three ways long-running agents handle context loss differently — blaizedsouza · 2026-07-28
- Kimi K3 scales Kimi Linear to 2.8T parameters and drops RoPE for NoPE — rasbt · 2026-07-28
- A closed-loop system ran 633 experiments to find a two-metal N2O catalyst — bravo_abad · 2026-07-28