Metadata-Version: 2.4
Name: windscangeo
Version: 2025.0
Summary: Easy extraction and processing of matching GOES geostationary satellite and polar orbiting scatterometer data for model training.
Project-URL: Homepage, https://windscangeo.github.io/
Author-email: Yongxing Loo  <yloo@tudelft.nl>
License-File: LICENCE
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Classifier: Programming Language :: Python :: 3
Requires-Python: >=3.8
Requires-Dist: cartopy
Requires-Dist: fsspec
Requires-Dist: h5netcdf
Requires-Dist: h5py
Requires-Dist: matplotlib
Requires-Dist: netcdf4
Requires-Dist: numpy
Requires-Dist: pandas
Requires-Dist: s3fs
Requires-Dist: scikit-learn
Requires-Dist: torch
Requires-Dist: tqdm
Requires-Dist: xarray
Description-Content-Type: text/markdown

# Windscangeo 

**WindScanGEO** is a deep learning framework developped to detect wind speeds from the images of the  *Geostationary Operational Environmental Satellite (GOES)* .
It is able to deliver high frequency wind speed predictions (every 10 minutes) at a 25x25km spacial resolution over the ocean. This is useful
for any research that looks into daily and small-scale spacial variability and evolution of wind speeds. It is based on a ResNet50
model that is trained daily on scatterometer data which allows inference of the model on the extent of an entire GOES image.

This framework is open-source and has been developped by Y.Loo, Dr.
J.Sun, Dr. G.George at the Geoscience and Remote Sensing department and
the Intelligent Systems department of the TU Delft.

The framework is distributed as a **Python package** that is easily usable to train and infer models at any time and location desired by the user. For more information, go to [windscangeo.github.io ](windscangeo.github.io)To directly use the package yourself, go to the *"Installation"* page.

For any questions or inquiries, please contact [yloo@tudelft.nl](mailto:yloo@tudelft.nl)
