Título:	Integration of vegetation inventory data and thematic mapper image for Amazonian successional and mature forest classification.
Autoria:	LU, D. BATISTELLA, M. MORAN, E. F.
Afiliação:	1: Indiana University-CIPEC; 2: Embrapa Monitoramento por Satélite; 3: Indiana University-ACT.
Ano de publicação:	2003
Referência:	In: ASPRS ANNUAL CONFERENCE, 2003, Anchorage, Alaska-EUA. Proceedings... [S.l.]: ASPRS, 2003.
Páginas:	13 p.
Conteúdo:	Successional and mature forest classification is often difficult in moist tropical regions. This paper explores vegetation stand structures of successional and mature forests and their spectral characteristics. Canonical discriminant analysis (CDA) was used to identify important stand parameters for secondary succession and mature forest classification. Correlation coefficient was used to analyze different stand parameter relationships and associated TM spectral signatures. Transformed divergence was used to analyze the separability of succession stages and mature forest based on the resultant images from CDA and principal component analysis (PCA), respectively. This study indicates that five vegetation categories, i.e., initial succession, intermediate succession, advanced succession, small biomass mature forest, and large biomass mature forest, can be istinguished based on vegetation stand features using field measurements, but some of them are difficult to be classified using TM data. Tree diameter at breast height, tree height, aboveground biomass, and ratio of tree biomass to total aboveground biomass are the best stand parameters distinguishing vegetation classes. Bands TM 4 and TM 5 are best for distinguishing vegetation classes. The transformation using CDA improved separability of vegetation classes, but not using PCA. Two successional stages and one mature forest class are suitable in this study area.
Thesagro:	Floresta Vegetação
Palavras-chave:	mapeamento Amazonas
Tipo do material:	Artigo em anais e proceedings
Acesso:	openAccess
Aparece nas coleções:	Artigo em anais de congresso (CNPM)