Sat3R: Satellite DSM Reconstruction via RPC-Aware Depth Fine-tuning

TL;DR

Sat3R is a fast, RPC-aware depth fine-tuning framework that adapts foundation models for accurate satellite DSM reconstruction, outperforming zero-shot baselines and rivaling optimization methods.

Contribution

Introduces Sat3R, a novel RPC-aware depth fine-tuning approach that enables foundation models to accurately reconstruct satellite DSMs without per-scene optimization.

Findings

Sat3R reduces MAE by 38% over zero-shot baselines.

Achieves competitive accuracy with optimization-based methods.

Provides over 300x speedup in DSM reconstruction.

Abstract

Accurate Digital Surface Model (DSM) reconstruction from satellite imagery is critical for applications such as disaster response, urban planning, and large-scale geographic mapping. Existing approaches face a fundamental trade-off: optimization-based methods achieve strong accuracy but require hours of per-scene computation, while generalizable geometry foundation models offer near-instant inference but fail to generalize to satellite imagery due to the domain gap introduced by the Rational Polynomial Camera (RPC) model and mismatched depth scale distributions. We present Sat3R, a feed-forward framework that bridges this gap via RPC-aware metric depth fine-tuning of Depth Anything V2 using the Scale-Invariant Logarithmic (SiLog) loss. By constructing physically consistent pseudo depth supervision from RPC geometry, Sat3R adapts a monocular depth foundation model to the satellite domain without per-scene optimization. Experiments on the DFC2019 benchmark demonstrate that Sat3R reduces MAE by 38% over zero-shot feed-forward baselines and achieves competitive accuracy against optimization-based methods, while delivering over 300x speedup. Sat3R demonstrates that feed-forward models, when properly adapted to the satellite domain, can match optimization-based accuracy at a fraction of the computational cost, paving the way for practical large-scale satellite DSM reconstruction.