Transformação por componentes principais e por decorrelação de bandas aplicadas à caracterização de feições da paisagem do Nordeste paraense.
As a support to the land use change studies in the Amazon, this work deals with the use of transformations of principal components and band decorrelation, for a spectral characterization features in the landscape of Igarapé-Açu, Pará, Brazil. A study module of 16 km x 13 km was selected from Landsat/TM image in order to apply the spectral rotation, considering six bands of reflective spectra the TM sensor. For a principal components analysis the three first images were used, while for a transformation of band decorrelation the TM 3-4-5 bands were considered , using the file obtained in spectral rotation process. Through products genereted by these techniques it was obtained photographic products that were qualitatively evaluated in the field and in the laboratory, in order to observe the agreement between enhanced spectral characteristics with vegetation cover and land use patterns observed on the field. It was noticed that, in general, the products obtained for both techniques have a signicative performance to detect and discriminate the features of interest. However, the color composition obtained for principal component analysis showed a better comparative performance, because there is a formation of pure spectral and strongly saturated colors to enhance land use features.
| Main Authors: | , , |
|---|---|
| Other Authors: | |
| Format: | Anais e Proceedings de eventos biblioteca |
| Language: | Portugues pt_BR |
| Published: |
In: SIMPÓSIO BRASILEIRO DE SENSORIAMENTO REMOTO, 8., 1996, Salvador. Anais... São Paulo: INPE: Sociedade Latino-Americana de Sensoriamento Remoto e Sistemas de Informações Espaciais, 1996.
2022-10-24
|
| Subjects: | Slash-and-Burn Agriculture, Image Enhancement, Land Use/Cover, Agricultura de corte e queima, Uso da Terra, Cobertura do Solo, |
| Online Access: | http://www.alice.cnptia.embrapa.br/alice/handle/doc/1147686 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|