Skip to content

Dataset · onboard-coop

Hyperspectral Satellite Images for Methane Search

onboard-coop/STARCOP-fast-products

Satellite data for methane emission segmentation, prepared for lightweight onboard neural networks.

What's inside

The dataset includes 1,000 to 10,000 geospatial images formed from 72 hyperspectral channels in the 2122-2488 nm wavelength range. The materials are prepared for training and evaluating models from a research paper on optimizing methane detection directly onboard satellites.

What you can do with it

  • Training semantic segmentation models to detect methane leaks from space
  • Benchmarking lightweight and fast neural networks for constrained onboard hardware
Tasks
Image segmentation
Size
1K–10K rows
Downloads this month
44.5K
Likes
1

How to load it

from datasets import load_dataset

ds = load_dataset("onboard-coop/STARCOP-fast-products")