GeoFormer: A Lightweight Swin Transformer for Joint Building Height and Footprint Estimation from Sentinel Imagery

TL;DR

GeoFormer is a lightweight Swin Transformer model that accurately estimates building height and footprint from Sentinel data, outperforming CNN baselines and demonstrating strong transferability across diverse cities.

Contribution

The paper introduces GeoFormer, a novel multi-task transformer framework that jointly estimates building parameters using open-access satellite data, with comprehensive evaluation and public release.

Findings

GeoFormer achieves a 3.19 m RMSE in building height estimation.

It outperforms CNN baselines like UNet by 7.5%.

A 5x5 receptive field and DEM data are optimal for height estimation.

Abstract

Building height (BH) and footprint (BF) are fundamental urban morphological parameters required by climate modelling, disaster-risk assessment, and population mapping, yet globally consistent data remain scarce. In this work, we develop GeoFormer, a lightweight Swin Transformer-based multi-task learning framework that jointly estimates BH and BF on a 100 m grid using only open-access Sentinel-1 SAR, Sentinel-2 multispectral, and DEM data. A geo-blocked data-splitting strategy enforces strict spatial independence between training and evaluation regions across 54 morphologically diverse cities. We set representative CNN baselines (ResNet, UNet, SENet) as benchmarks and thoroughly evaluate GeoFormer's prediction accuracy, computational efficiency, and spatial transferability. Results show that GeoFormer achieves a BH RMSE of 3.19 m with only 0.32 M parameters -- outperforming the best CNN baseline (UNet) by 7.5% -- indicating that windowed local attention is more effective than convolution for scene-level building-parameter retrieval. Systematic ablation on context window size, model capacity, and input modality further reveals that a 5x5 (500 m) receptive field is optimal, DEM is indispensable for height estimation, and multispectral reflectance carries the dominant predictive signal. Cross-continent transfer tests confirm BH RMSE below 3.5 m without region-specific fine-tuning. All code, model weights, and the resulting global product are publicly released.