Iterative Mapping of Probabilities



This repository contains data and scripts for implementing the Iterative Mapping of Probabilities (IMP) algorithm proposed in the preprint submitted to the International Journal of Applied Earth Observation and Geoinformation (JAG). The framework aims to improve the accuracy of land cover mapping by iteratively refining class maps to match independent area statistics. The experiment focuses on generating classification maps for five countries (Belgium, Czechia, Germany, Luxembourg, Netherlands) based on input probability rasters.
Date made available9 Feb 2024
PublisherWageningen University & Research


  • land cover
  • uncertainty
  • post-processing
  • area statistics

Cite this