Anais do XX Simpósio Brasileiro de Sensoriamento Remoto

Our goal in this study was to perform a LULC classification for the southern part of Roraima state. This area has a highly frequent cloud cover and a lack of LULC information. We used a SAR-Optical multisensory methodology, with a cloud computing process, to be able to classify all the areas, with less computational effort and in less time. Our results show an Overall Accuracy of 92.61%, with Users' and Producers' Accuracy (UA and PA), around 90% for all ten classes. Also, this approach identified important classes for the region, such as perennial crops and conversion areas.