Hugin EO: Machine Learning for Earth Observation made easy

Hugin helps scientists run Machine Learning experiments on geospatial raster data.

Hugin was developed as part of the ESA funded ML4EO

Overall Hugin aims to facilitate experimentation with multiple machine learning problems, like:

• Classification
• Segmentation
• Super-Resolution

Currently Hugin builds on top of the Keras machine learning library but it also aims to support, in the future, additional backends like scikit-learn.

• Installation
  • Prerequisites
  • Using pip
    • From PyPi
    • From GitHub
  • From source code
• Introduction
  • Standalone
    • Training
      • Global Configuration
      • Data Source Specification
      • Model Configuration
      • Example Experiment
    • Prediction
    • Mapping

Indices and Tables