Favoritar este trabalho

The Amazon region, by its geographical location, with influences of the Intertropical Convergence Zone, offers to optical sensors, the obtainment of images with high cloud coverage in the greater part of the year, which makes difficult and/or prevents the orbital images’ interpretation and can become an obstacle for the remote sensing and geoprocessing in the forests’ monitoring. Keeping this in view, it was implemented a cloud automatic detection tool, the Cloud Detection, which is available in the application TerraAmazon, developed by INPE and FUNCATE to give support to the projects, developed by the National Institute for Space Research (INPE) for the monitoring of forests. The present work applied different values to the parameters referring to the plugin Cloud Detection in the semiautomatic detection of clouds and shadows for WFI/CBERS-4 images, to verify which parameters collaborate for the efficiency of the detection and vectorization of the two targets in question. The results showed that the values of the morphological filter’s opening was the one who collaborates more for the differences in the vectorization; and there will always have confusion related to some water bodies that, because of theirs similar reflectance, they are detected as shadows, which can be edited manually and deleted of the mapping.