ABSTRACT
This study evaluated the result of different methodologies in determining the use and occupancy of land using digital aerial photographs with spatial resolution of 1 m (scenario 01) and satellite imagery with a spatial resolution of 0.5 m (scenario 02), photo-interpreted screen, on a scale of 1:2.000 mapping, watershed 'Horizonte' stream, in Alegre, ES state. We used the global performance indices and Kappa index for determining the average error between the sets and t test at 5 % probability to assess the degree of significance in the process of image interpretation between scenarios 01 and 02. We identified 27 classes of land use for scenarios 01 and 02. The classes of pasture (599.62 and 442.43 ha) and forest fragmentation (319.45 and 258.07 ha) respectively, occupying the largest areas in the two scenarios, composing 69.80 % for scenario 01 and 53.04 % for scenario 02 of the total area. There was a positive variation of 6.67 % in the DG scenario in scenarios 01 and 02. However, the K shows that the scenario 02 has a lower rate of errors of omission and commission of 2.09 %, but does not present significance by the statistical analysis presented at 5 % level by t test.
Keywords:
environmental conservation; remote sensing; geographic information systems