This package uses machine learning for the land cover classification of Sentinel-1 data. It is the python programming project of Theresa Möller and Antje Uhde for the module Geo419 of the M.Sc. Geoinformatics course at the Friedrich-Schiller-University Jena.
Please note that it is recommended to use the package with Python 3.7.
To use this package, download or clone the repo. In a cmd shell move to
the folder where setup.py is located and type python setup.py sdist bdist_wheel
(make sure to have setuptools wheel installed, e.g. via pip). Type
pip install dist/RanForCorine-VERSION-py3-none-any.whl (replace VERSION
with the correct version number) in the cmd shell. GDAL
needs to be installed separately due to a more complex installation process.
You can download the wheel here.
Chose the correct version specified in the requirements.txt which fits
your operating system. Use pip install gdalfilename.whl.
Make sure to update of the system environment variables
(check step 3 of this description).
Due it its dependency, the same accounts for rasterio, which can simply be
installed using pip.
You should be good to go now!
If you want to use RanForCorine with in an Anaconda environment there are some additional steps to follow.
First, move to the folder ranforcorine-conda and run conda-build . from an anaconda powershell.
After that you run conda install --use-local RanForCorine to finally install the package.
See the jupyter notebook in /examples for further information on how to use this package.