ZoomEarth: Active Perception for Ultra-High-Resolution Geospatial Vision-Language Tasks

TL;DR

ZoomEarth introduces an active perception framework for ultra-high-resolution remote sensing images, enabling models to selectively revisit regions, improving task performance and versatility across various geospatial vision-language tasks.

Contribution

The paper presents ZoomEarth, a novel adaptive cropping-zooming framework with a new Region-Guided reward, and a large-scale UHR remote sensing benchmark dataset, LRS-GRO, advancing active perception in this domain.

Findings

ZoomEarth achieves state-of-the-art results on LRS-GRO and public benchmarks.

The framework effectively guides models to focus on information-rich regions.

ZoomEarth demonstrates strong versatility across downstream geospatial tasks.

Abstract

Ultra-high-resolution (UHR) remote sensing (RS) images offer rich fine-grained information but also present challenges in effective processing. Existing dynamic resolution and token pruning methods are constrained by a passive perception paradigm, suffering from increased redundancy when obtaining finer visual inputs. In this work, we explore a new active perception paradigm that enables models to revisit information-rich regions. First, we present LRS-GRO, a large-scale benchmark dataset tailored for active perception in UHR RS processing, encompassing 17 question types across global, region, and object levels, annotated via a semi-automatic pipeline. Building on LRS-GRO, we propose ZoomEarth, an adaptive cropping-zooming framework with a novel Region-Guided reward that provides fine-grained guidance. Trained via supervised fine-tuning (SFT) and Group Relative Policy Optimization (GRPO), ZoomEarth achieves state-of-the-art performance on LRS-GRO and, in the zero-shot setting, on three public UHR remote sensing benchmarks. Furthermore, ZoomEarth can be seamlessly integrated with downstream models for tasks such as cloud removal, denoising, segmentation, and image editing through simple tool interfaces, demonstrating strong versatility and extensibility.