This repository curates Visual Language Models (VLMs) and associated resources specifically for Earth Observation (EO). It serves researchers and practitioners by consolidating papers, code, and datasets for tasks like image captioning, text-image retrieval, visual grounding, and visual question answering within the remote sensing domain. The project aims to accelerate VLM adoption in EO by providing a centralized, organized overview of the state-of-the-art.

How It Works

The project functions as a structured, curated index of VLM research relevant to Earth Observation. It categorizes resources by task, including foundation models, image captioning, text-image retrieval, visual grounding, and visual question answering. For each entry, it typically links to the research paper and, where available, the corresponding code repository. This approach highlights advancements and facilitates discovery of specialized EO VLM solutions, bridging the gap between general VLM capabilities and the unique challenges of remote sensing data.

Quick Start & Requirements

This repository is a curated list and does not involve direct installation or execution. Users are directed to individual research papers and code repositories for specific setup, dependencies (e.g., GPU, CUDA, Python versions), and execution instructions. Links to datasets are also provided, with their own specific requirements.

Highlighted Details

• Comprehensive coverage of key VLM tasks tailored for remote sensing imagery.
• Extensive listing of specialized Earth Observation datasets for VLM applications.
• Direct links to code repositories for many listed research papers, promoting reproducibility.
• Strong emphasis on recent (2023-2024) advancements, reflecting an active research field.

Maintenance & Community

The list is maintained by the GEOspatial Artificial Intelligence (GEOAI) research group at the National Center for Remote Sensing - CNRS, Lebanon. Contributions are encouraged. No specific community channels (e.g., Discord, Slack) are listed.

Licensing & Compatibility

As a curated list, this repository does not have a specific software license. The licensing and compatibility of individual linked papers, code repositories, and datasets will vary and must be assessed independently by the user.

Limitations & Caveats

This is an informational resource, not a deployable tool; users must evaluate and integrate individual VLM projects. The scope is limited to what is curated, and links may become outdated. No direct support or unified API is provided.