GeoRoPE: Ground-Aware Rotary Adaptation for Remote Sensing Foundation Models

2026-06-08

AI Summary

This work addresses the challenge of scale mismatch in downstream tasks when applying remote sensing foundation models pretrained across diverse sensors and ground sampling distances (GSDs), which often leads to physically inconsistent positional priors and poor adaptation to varying spatial granularities. To overcome this, the authors propose GeoRoPE, a lightweight adapter-based approach that enables efficient, spatially aware fine-tuning by freezing the original positional encodings while introducing geographically informed corrections. The core innovation lies in jointly integrating Geographic Coordinate Calibration (GCC) and Geographic Frequency Calibration (GFC) to construct GSD-invariant relative coordinate representations and scene-adaptive, position-sensitive modeling. Experiments demonstrate that GeoRoPE substantially enhances cross-scale robustness and representational capacity of various remote sensing foundation models across heterogeneous sensors, resolutions, and downstream tasks.

Abstract

Remote-sensing foundation models (RSFMs) benefit from pretraining on imagery from multiple sensors and ground sampling distances (GSDs), but such exposure alone does not resolve scale mismatch during downstream adaptation. A fixed token-grid offset can correspond to different ground distances across sensors, making grid-based positional priors physically inconsistent. Meanwhile, heterogeneous spatial granularity means that compact urban regions and homogeneous landscapes may require different positional sensitivities even under the same GSD. Therefore, we propose {GeoRoPE}, a ground-aware, RoPE-compatible, and parameter-efficient spatial adaptation method for RSFMs. GeoRoPE recalibrates token-level positional interactions from two complementary aspects. First, \textit{Geo-Coordinate Calibration (GCC)} rescales raw token-grid offsets according to the ground distance represented by one token-grid step, producing geo-calibrated relative coordinates across GSDs. Second, \textit{Geo-Frequency Calibration (GFC)} adjusts the native RoPE frequency with a relation-specific factor, enabling position sensitive adaptation to scene-dependent spatial granularity. GeoRoPE is injected into pretrained RSFMs through a lightweight adapter, preserving the frozen spatial prior while adding geo-aware positional corrections. Experiments across multiple RSFMs, sensors, resolutions, and downstream tasks demonstrate that GeoRoPE improves cross-resolution robustness and scale-sensitive representation learning.

Problem

• scale mismatch
• positional inconsistency
• spatial granularity
• ground sampling distance
• remote sensing foundation models

Innovation

• GeoRoPE
• ground-aware adaptation
• rotary position embedding
• cross-resolution robustness
• spatial granularity