EO-learn is an open-source library bridging the gap between EO data and existing ML technologies. A collection of Python packages has been developed to seamlessly access and process spatio-temporal image sequences acquired by any satellite fleet in a timely and automatic manner.
The design of eo-learn is modular and encourages collaboration – sharing and re-using specific tasks in a typical EO-value-extraction workflow, such as cloud masking, image co-registration, feature extraction, classification, etc. Since it is open source, everyone is free to use any of the available tasks and is encouraged to improve the existing tasks, develop new ones and share them with the rest of the community. A collection of Jupyter notebook examples was prepared, showcasing how to couple eo-learn with the most popular ML frameworks, such as scikit-learn, Keras, PyTorch and fastai.
These examples allow users to quickly set up their environment using tools such as Docker and AWS SageMaker to ease the training and prediction procedure. Several AI-ready datasets are available, such as automatic label retrieval from OpenStreetMap, high-resolution imagery from Mapbox and medium resolution from Sentinel, urban settlements in Europe, land cover datasets as well as many other open-source datasets.
Sample use-cases were demonstrated, readily available for reuse, covering water monitoring on a global scale, land cover segmentation, mapping of urban settlements and others.
Share