This study aimed to estimate and map areas planted with soybean [Glycine max (L.) Merr.] through multitemporal images EVI/MODIS and classification of images based on geo-object. The study area comprised the southern part of the State of Maranhão. For the mapping of the soybean crop the Enhanced Vegetation Index (EVI) and the Crop Enhancement Index (CEI) for image classification sensor-system Terra/MODIS was used. For this calculation twelve images were used, including offseason and harvest of the crop, as per the state agricultural calendar. In addition, the segmentation was employed using scaling parameters 250, the algorithms "classification" and "merge region", and extracting attributes for classification GEOgraphic-Object-Based Image Analysis (GEOBIA). To assess the accuracy of the classification, parameters Kappa and Accuracy Global and its resulting Z test were applied. A null hypothesis (H0) of equal and opposite rates for their differences (H1) at 0.05 level of significance was established. The results indicate that the proposed methodology is efficient for mapping the soybean crop, with 0.89 for the parameter Kappa.
digital image classification; crop area estimates; image segmentation; remote sensing
