ZoomEarth：面向超高分辨率地理空间视觉语言任务的主动感知
ZoomEarth: Active Perception for Ultra-High-Resolution Geospatial Vision-Language Tasks
把「主动感知」引入超高分辨率视觉语言模型：模型通过自适应裁剪-缩放反复回看信息密集区域，并用区域引导奖励配合 GRPO 训练，可直接迁移到任意需要细粒度小目标的高分辨率 VLM 任务（遥感、病理、文档、卫星图）。
原文依据 we propose ZoomEarth, an adaptive cropping-zooming framework with a novel Region-Guided reward that provides fine-grained guidance
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论超高分辨率（UHR）遥感（RS）图像蕴含丰富的细粒度信息，但也给有效处理带来了挑战。现有的动态分辨率与词元剪枝方法受限于被动感知范式，在获取更精细视觉输入时反而遭受冗余增加的问题。在本工作中，我们探索了一种新的主动感知范式，使模型能够重新回看信息丰富的区域。首先，我们提出 LRS-GRO，一个专为 UHR 遥感处理中的主动感知而构建的大规模基准数据集，涵盖全局、区域和对象三个层次的 17 种问题类型，并通过半自动流程完成标注。基于 LRS-GRO，我们提出 ZoomEarth，一个自适应裁剪-缩放框架，并设计了新颖的区域引导奖励以提供细粒度指导。通过监督微调（SFT）与组相对策略优化（GRPO）训练，ZoomEarth 在 LRS-GRO 上取得最优性能，并在零样本设置下于三个公开 UHR 遥感基准上取得最优性能。此外，ZoomEarth 可通过简单的工具接口无缝集成到去云、去噪、分割和图像编辑等下游模型，展现出很强的通用性与可扩展性。所有数据与代码将发布于 https://earth-insights.github.io/ZoomEarth。
英文原文摘要 1355 字符
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. All data and code will be released at https://earth-insights.github.io/ZoomEarth.