SkyScript: A Large and Semantically Diverse Vision-Language Dataset for Remote Sensing

TL;DR

SkyScript is a large, diverse remote sensing vision-language dataset created using geo-coordinates and OpenStreetMap, enabling significant improvements in zero-shot scene classification and other multi-modal remote sensing tasks.

Contribution

The paper introduces SkyScript, a novel large-scale remote sensing dataset linking images with rich semantic tags via geo-coordinates, facilitating the development of versatile vision-language models.

Findings

Achieved 6.2% accuracy improvement in zero-shot scene classification

Demonstrated effective zero-shot transfer for object attribute classification

Enabled cross-modal retrieval and captioning in remote sensing

Abstract

Remote sensing imagery, despite its broad applications in helping achieve Sustainable Development Goals and tackle climate change, has not yet benefited from the recent advancements of versatile, task-agnostic vision language models (VLMs). A key reason is that the large-scale, semantically diverse image-text dataset required for developing VLMs is still absent for remote sensing images. Unlike natural images, remote sensing images and their associated text descriptions cannot be efficiently collected from the public Internet at scale. In this work, we bridge this gap by using geo-coordinates to automatically connect open, unlabeled remote sensing images with rich semantics covered in OpenStreetMap, and thus construct SkyScript, a comprehensive vision-language dataset for remote sensing images, comprising 2.6 million image-text pairs covering 29K distinct semantic tags. With continual pre-training on this dataset, we obtain a VLM that surpasses baseline models with a 6.2% average accuracy gain in zero-shot scene classification across seven benchmark datasets. It also demonstrates the ability of zero-shot transfer for fine-grained object attribute classification and cross-modal retrieval. We hope this dataset can support the advancement of VLMs for various multi-modal tasks in remote sensing, such as open-vocabulary classification, retrieval, captioning, and text-to-image synthesis.