This work aimed at to recognize and to analyze the changes in the landscape of the Pacaraima Mountain that is located in the extreme north of the Roraima State, Brazil. For that, four images of the satellite LANDSAT-5/TM were used referring the dates of covering 17/02/1990, 04/04/1995, 17/05/1999 and 06/05/2004, on which was determined a study area, and one radar image of the project SRTM, to characterize the regional relief. In the LANDSAT-5/TM images processing techniques were applied to minimize the current mistakes of your acquisition. The segmentation technique by region growing was employ, together at field work, in order to recognize the units of local landscape, and the technique of supervised classification MAXVER, with the intention of to estimate the representation of landscape units along the multitemporal series. In SRTM image the processing consisted in the analysis of MNT and generation of hipsometric and slope maps. This way, four units of landscape were recognized and characterized as: landscape unit of forest (UPF); savanna (UPS); deforestation (UPD); and human occupation (UPOH).
|
