Beyond GSD-as-Token: Continuous Scale Conditioning for Remote Sensing VLMs

TL;DR

This paper introduces ScaleEarth, a novel framework for remote sensing vision-language models that dynamically conditions on ground sampling distance (GSD) as a continuous variable, improving performance across diverse Earth-system tasks.

Contribution

The paper proposes CS-HLoRA, a continuous scale-conditioning method, and a new scale-aware dataset, GeoScale-VQA, advancing remote sensing VLMs beyond discrete GSD tokens.

Findings

ScaleEarth achieves state-of-the-art results on remote sensing benchmarks.

CS-HLoRA effectively modulates model computation based on GSD.

The approach enables dynamic routing and GSD prediction without sensor metadata.

Abstract

Remote sensing vision-language models (RS-VLMs) face a fundamental mismatch with natural-image counterparts: the same geographic object exhibits radically different visual evidence across ground sampling distances (GSDs) spanning multiple orders of magnitude. Yet existing RS-VLMs often discard GSD or inject it as a discrete text token, forcing a single static parameter set to absorb the entire scale spectrum. We introduce ScaleEarth, a parameter-efficient fine-tuning framework built on Qwen3-VL that treats GSD as a continuous conditioning variable governing the model's computation path. At its core, CS-HLoRA (Continuous Scale-Conditioned Hyper-LoRA) modulates the LoRA low-rank subspace through a GSD-driven gate, enabling the model to dynamically route computation by physical scale. To remove reliance on sensor metadata at deployment, we pair CS-HLoRA with SSE-U, a lightweight heteroscedastic sub-head that predicts GSD and its uncertainty from visual features. To provide matching supervision, we construct GeoScale-VQA, a 1.5M-sample scale-layered RS-VQA corpus whose question-answer generation is conditioned on the same physical scalar that drives CS-HLoRA, forming a closed method-data loop. Trained with QLoRA on an 8B backbone, ScaleEarth achieves state-of-the-art results on remote-sensing benchmarks covering diverse Earth-system tasks, including XLRS-Bench and OmniEarth-Bench.