LandCoverPy, a scalable land cover/land use classification workflow

A scalable land cover classification workflow aimed to be able to scale to cover the Mediterranean bassin.

A research article describing the methodology followed on this workflow can be found at:

Scalable approach for high-resolution land cover: a case study in the Mediterranean Basin.

Antonio Manuel Burgueño, José F. Aldana-Martín, María Vázquez-Pendón, Cristóbal Barba-González, Yaiza Jiménez Gómez, Virginia García Millán & Ismael Navas-Delgado

Journal of Big Data 10, 91 (2023). doi: 10.1186/s40537-023-00770-z

Installation

The package is not available on PyPI, so you need to install it from the source code. To do so, you can clone the repository and install it using pip:

git clone https://github.com/KhaosResearch/landcoverpy.git
cd landcoverpy
pip install .

For development purposes, you can install the package in editable mode:

pip install -e .

Usage

An usage example can be found at the main usage notebook. Example input data in different formates can be found at validated_data. Example label mappings can be found at label_mappings.

Environment variables

Most of the configuration is done through environment variables. You should create a .env file following the .env.template file. Default values can be observed in config.py.

License

This project is licensed under the MIT license. See the LICENSE file for more info.