geo2ml
Python library and module for converting earth observation data to be suitable for machine learning models
- Sampling features from a raster using point geometries or polygons
- Tiling larger rasters and shapefiles into smaller patches
- Rasterizing polygon geometries for semantic segmentation tasks
- Converting vector data to COCO and YOLO formats and creating required dataset files
- Visualization of generated datasets
Install
First install GDAL to your system. If you use conda then installing rasterio
is enough, but with pip use instructions from https://pypi.org/project/GDAL/.
Then you can install the package by
pip install git+https://github.com/mayrajeo/geo2ml.git
If you want to have an editable install then first clone the repository
git clone https://github.com/mayrajeo/geo2ml.git
cd geo2ml
pip install -e .
How to use
Running geo2ml_help
shows all available commands. Documentation for each command is found by running for example geo2ml_sample_points --help