Mapping land use and cover in the southeastern Iberian Peninsula in 2000-2004 using Landsat imagery

This paper explains the methodology used to produce land cover mapping for the east of Andalusia (29.259 km²) in 2000-2004. We used a hybrid classifier using Landsat imagery and auxiliary variables. Training areas were obtained semi-automatically from the SIOSE database and filtered with NDVI thresholds. Testbeds were conducted based on the variables in cluded i n th e st atistical cla ssifier. Th e b est re sults in terms of o verall a ccuracy and cla ssified a rea were obtained by e xcluding w inter s olar r adiation and band 1. T he pr ocess w as c arried o ut sep arately f or natural and urbanized areas, with an overall accuracy of over 88%, and for crops with an accuracy of over 86%.