SatGeo-NeRF: Geometrically Regularized NeRF for Satellite Imagery

XXV ISPRS CONGRESS 2026
Valentin Wagner 1, Sebastian Bullinger 1, Michael Arens 1, and Rainer Stiefelhagen 2,
1 Fraunhofer Institute of Optronics, System Technologies and Image Exploitation
2 Karlsruhe Institute of Technology
Paper arXiv

Abstract

We present SatGeo-NeRF, a geometrically regularized NeRF for satellite imagery that mitigates overfitting-induced geometric artifacts observed in current state-of-the-art models using three model-agnostic regularizers. Gravity-Aligned Planarity Regularization aligns depth-inferred, approximated surface normals with the gravity axis to promote local planarity, coupling adjacent rays via a corresponding surface approximation to facilitate cross-ray gradient flow. Granularity Regularization enforces a coarse-to-fine geometry-learning scheme, and Depth-Supervised Regularization stabilizes early training for improved geometric accuracy. On the DFC2019 satellite reconstruction benchmark, SatGeo-NeRF improves the Mean Altitude Error by 14.0% and 11.4% relative to state-of-the-art baselines such as EO-NeRF and EO-GS.

Overview

We augment a satellite-domain–adapted backbone (EO-NeRF, Marı́ et al. (2023)) with three model-agnostic regularizers: (1) Gravity-Aligned Planarity Regularization: estimate explicit normals by sampling two adjacent rays per training ray and evaluating depth only; align normals with the axis of gravity to mitigate overfitting and enable cross-ray smoothing. (2) Geometrical Granularity Regularization: coarse-to-fine masking of the input frequency encoding. (3) Depth-Supervised Regularization: depth constraints from a sparse 3D point cloud obtained from a Bundle Adjustment preprocessing step.

Gravity-Aligned Planarity Regularization: A local plane is estimated from three surface points on adjacent rays; its normal is constrained to align with the axis of gravity (approximated as UTM-up).

Results

EO-NeRF
Ours

JAX 004

EO-NeRF
Ours

JAX 068

EO-NeRF
Ours

JAX 214

EO-NeRF
Ours

JAX 260

BibTeX


@misc{wagner2026satgeonerf,
  title={SatGeo-NeRF: Geometrically Regularized NeRF for Satellite Imagery}, 
  author={Valentin Wagner and Sebastian Bullinger and Michael Arens and Rainer Stiefelhagen},
  year={2026},
  eprint={2603.21931},
  archivePrefix={arXiv},
  primaryClass={cs.CV},
  url={https://arxiv.org/abs/2603.21931}, 
}