SatGeo-NeRF: Geometrically Regularized NeRF for Satellite Imagery

TL;DR

SatGeo-NeRF introduces a geometrically regularized NeRF model for satellite imagery that reduces artifacts and improves geometric accuracy through three novel regularizers, achieving better performance on a satellite reconstruction benchmark.

Contribution

It proposes three model-agnostic regularizers for NeRF to enhance geometric fidelity in satellite imagery reconstruction, addressing overfitting artifacts.

Findings

Improves Mean Altitude Error by 13.9% over EO-NeRF.

Enhances geometric accuracy on DFC2019 benchmark.

Reduces overfitting-induced artifacts in satellite NeRF models.

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 13.9% and 11.7% relative to state-of-the-art baselines such as EO-NeRF and EO-GS.