 This study assessed the accuracy of five different approaches for producing crop type maps from high resolution satellite imagery. The results showed that a random forest classifier was able to achieve overall accuracies above 80% for most sites, with only two sites showing lower performance. These sites were Madagascar due to the presence of small fields, and Burkina Faso due to a mix of trees and crops in the fields. The approach is based on supervised machine learning techniques, which require in situ data collection for the training step, but the map production is fully automated. This article was authored by Jordi Inglada, Marcella Arias, Benjamin Tardy, and others.