Cartografia de Coberto do Solo para o Território Angolano Utilizando Imagens de Satélite Modis

Abstract

The objective of this work is to obtain a land cover map of Angola, using daily images of four months of 2003, from the Moderate Resolution Imaging Spetroradiometer (MODIS) sensor. Since those images are subject to atmospheric contamination and to the existence of clouds, it is necessary to apply a multitemporal composite criterion to remove these effects. The application of this criterion produces monthly composite images that are used in a classification algorithm tree. The selection of image data for training the classifier and for accuracy assessment was supported by maps at different scales, vegetation books, expert knowledge, and Landsat Thematic Mapper imagery. The classification accuracy was based on crossvalidation method, which uses the same dataset in the validation and in the classification and in a method based on an independent dataset. The result is an up-to-date and uniform map for 2003, with a method which allow to be used to produce maps of different years in order to detect land cover changes.