BDBComp
Parceria:
SBC
MAPEAMENTO, CLASSIFICAÇÃO E DIMENSIONAMENTO DE ÁREAS GEOGRÁFICAS POR SENSORIAMENTO REMOTO UTILIZANDO PROCESSAMENTO DE IMAGENS

Paulo Domingues de Oliveira JuniorMarcelo Andrade da Costa Vieira

The aim of this work is to develop image processing techniques to analyze satellites imagesacquired for remote sensing, in order to classify different types of geographic areas from an automatic methodology of segmentation. The main objective consists in developing a low cost computer program in order to detect and classify geographic areas of interest from images obtained by C-BERS2 satellite. For each geographic area the satellite gives a set of multi spectral images (different channels), which one corresponding to the detected wavelength band. Subsequently, overlapping and comparison methods from these multi spectral images were developed to classify the proposed zones. At the beginning, it was identified areas where its characteristics could be classified using only one channel (one image). After this, it was identified the zones that needs to be analyzed using more than one channel for identifications and classification, besides there were ones that do not have well defined standards. The first step, realized by using only some channels recommended by INPE, do not achieved satisfactory results. However, using different bands and a large number of images, the classification became more precise. The developed applicative is able to mark and classify automatically the detected zones in the images obtained by C-BERS2 using standard colors and gives information about areas of each region found.

http://www.lbd.dcc.ufmg.br/colecoes/wvc/2006/0032.pdf

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato: bdbcomp@lbd.dcc.ufmg.br
     Mantida por:
LBD