The objective of this work was to assess sugarcane area mapping using six‑year‑old time series of normalized difference vegetation index (NDVI) data from the Vegetation sensor on board of the "système pour l'observation de la Terre" (SPOT) satellite. Three land cover classes (sugarcane, pasture, and forest), from the state of São Paulo, Brazil, were selected as reference spectral‑temporal signatures, which were used as endmembers for spectral classification with the algorithm spectral angle mapper (SAM). Based on this classification, sugarcane areas were mapped by applying thresholds on the sugarcane rule image from the SAM, generated from values from the reference spectra. Results show that the SAM algorithm can be applied to time series of moderate resolution multi‑temporal data, allowing for the efficient mapping of an agricultural crop in a mesoregional scale. Official data from sugarcane areas for the microregions of the state of São Paulo are well correlated (r² = 0.8) with the data obtained with the evaluated method. SAM is a useful algorithm for time series analysis. NDVI time series from the SPOT Vegetation sensor can be used for low-resolution sugarcane area mapping.
temporal analysis; linear spectral mixture model; crop forecasting; agricultural remote sensing
