UnB - Instituto de Geociências - Produção Científica
RESUMOS/ABSTRACTS de artigos publicados em revistas com corpo editorial

Desenvolvimento de um Procedimento Simplificado do Método Multiple Endmember Spectral Mixture Analysis (MESMA) para o Tratamento de Imagens Hiperespectrais. 

Carvalho Jr,O.A.
Guimarães,R.F.
Meneses,P.R. 

In: Simpósio Brasileiro de Sensoriamento Remoto, 10, 2001, Foz do Iguaçu-PR, INPE. Anais(CD-ROM).

Abstract

For the analysis of spectral mixture the method of Linear Spectral Unmixing presents many mistakes because it is hard to adjust the endmembers existent in the image with the elements really existent in each pixels. A method proposed to eliminate that source of mistakes is the Multiple Endmember Spectral Mixture Models (MESMA) that defines for the root mean square which is the best association of spectral mixture that describes each pixel. This way, a routine was developed in IDL, which accomplishes that comparison of the models of mixtures defined for each pixel the most appropriate. This method was tested in the Niquelândia mine and showed excellent results being coherent with the field rising.