This repository serves as a comprehensive, curated collection of resources for Remote Sensing Foundation Models (RSFMs). It targets researchers and practitioners in the field of Earth observation, providing a centralized hub for papers, datasets, benchmarks, code, and pre-trained weights to accelerate the development and adoption of advanced AI models for remote sensing data.

How It Works

The repository categorizes RSFMs into several key areas: Vision, Vision-Language, Generative, Vision-Location, Vision-Audio, Task-specific, and Agents. It meticulously lists numerous models within each category, detailing their publication, associated papers, and available code/weights. This structured approach allows users to easily discover and compare state-of-the-art RSFMs.

Quick Start & Requirements

This repository is a curated list and does not have a direct installation or execution command. Users are directed to individual project links for specific setup instructions and requirements.

Highlighted Details

• Extensive catalog of over 100 Remote Sensing Foundation Models, categorized by modality and task.
• Includes a wide array of datasets and benchmarks specifically designed for evaluating RSFMs.
• Features a dedicated section for survey and commentary papers, offering insights into the field's progress and future directions.
• Provides links to relevant projects and a playground for evaluating and fine-tuning RSFMs.

Maintenance & Community

The repository was last updated on April 9, 2025, indicating recent activity. It includes citation information for key papers, suggesting community engagement and recognition.

Licensing & Compatibility

The repository itself is a collection of links and does not impose a specific license. Individual projects linked within the repository will have their own licensing terms, which users must consult.

Limitations & Caveats

Some entries indicate "Coming soon" for datasets or benchmarks, suggesting that certain resources are still under development or not yet publicly released. The repository is a curated list, not a unified framework, meaning users must navigate to individual project pages for practical implementation.