 This paper presents a methodology for producing accurate land cover maps using high-resolution optical image time series. The approach is based on supervised classification and uses existing databases as reference data for training and validation. The originality of the approach lies in the use of all available image data, simple pre-processing steps, and a robust supervised classifier. The produced maps have a CAPA coefficient of 0.86 with 17 land cover classes and are efficient, fast, and do not require expensive field surveys or human operators for decision-making. The land cover maps are provided with a confidence map that gives information at the pixel level about the expected quality of the result. This article was authored by Jordan Glotta, Arthur Vincent, Marcella Arias, and others.