chagmgang/dinov2-remote-sensing

Implementation dino v2 for remote sensing with huggingface transformers

23
/ 100
Experimental

This project offers pre-trained vision models specifically designed for analyzing satellite and aerial imagery. It takes raw remote sensing images and produces highly accurate classifications and features, which can then be used for tasks like land-use mapping, environmental monitoring, or urban planning. Earth observation specialists, geospatial analysts, and researchers working with satellite data would find this useful for developing advanced image analysis applications.

No commits in the last 6 months.

Use this if you need to extract meaningful features or classify scenes from large volumes of remote sensing images, especially when training a model from scratch is not feasible.

Not ideal if you are working with standard photographic images or require real-time processing on embedded devices with limited computational resources.

remote-sensing geospatial-analysis satellite-imagery land-cover-classification environmental-monitoring
No License Stale 6m No Package No Dependents
Maintenance 2 / 25
Adoption 7 / 25
Maturity 8 / 25
Community 6 / 25

How are scores calculated?

Stars

36

Forks

2

Language

Jupyter Notebook

License

Last pushed

Jul 30, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/transformers/chagmgang/dinov2-remote-sensing"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.