@MastersThesis{Xaud:1998:AvDaTM,
author = "Xaud, Maristela Ramalho",
title = "Avalia{\c{c}}{\~a}o de dados TM/LANDSAT e SAR/JERS na
caracteriza{\c{c}}{\~a}o da cobertura vegetal e
distribui{\c{c}}{\~a}o de fitomassa em {\'a}reas de contato
floresta/savana no Estado de Roraima - Brasil",
school = "Instituto Nacional de Pesquisas Espaciais (INPE)",
year = "1998",
address = "S{\~a}o Jos{\'e} dos Campos",
month = "1998-04-23",
keywords = "x.",
abstract = "OBS.: est{\~a}o faltando as p{\'a}ginas 24, 58, 130 e 136 no
original. A necessidade de quantifica{\c{c}}{\~a}o da cobertura
vegetal decorre da demanda de dados para estudos relacionados com
os processos de mudan{\c{c}}as globais. O presente estudo
apresenta par{\^a}metros biof{\'{\i}}sicos da cobertura vegetal
em {\'a}reas de contato de florestas e savanas no Estado de
Roraima-Brasil, avaliando suas rela{\c{c}}{\~o}es com dados
orbitais (TM/LANDSAT e SARAERS-1). As comunidades vegetais
presentes na {\'a}rea de estudo foram descritas segundo as
caracter{\'{\i}}sticas fision{\^o}mico-estruturais e de
composi{\c{c}}{\~a}o flor{\'{\i}}stica, classificando-se em
floresta prim{\'a}ria, sucess{\~a}o secund{\'a}ria, savana
arborizada, savana parque e savana gram{\'{\i}}neo-lenhosa.
Atrav{\'e}s da utiliza{\c{c}}{\~a}o de t{\'e}cnicas de
extra{\c{c}}{\~a}o de informa{\c{c}}{\~o}es sensoriadas, os
produtos originados nas faixas {\'o}ptica e de microondas foram
processados e os valores digitais foram extra{\'{\i}}dos. Para a
imagem do sensor TM, foi aplicada a t{\'e}cnica de Modelo Linear
de Mistura Espectral, gerando as imagens-propor{\c{c}}{\~o}es
(VEG, SOL e SOM), cujos valores digitais foram utilizados nas
rela{\c{c}}{\~o}es com os par{\^a}metros biof{\'{\i}}sicos
das diferentes fei{\c{c}}{\~o}es presentes na {\'a}rea de
contato. Com as imagens JERS e TM georreferenciadas, foi
poss{\'{\i}}vel a extra{\c{c}}{\~a}o dos valores digitais na
imagem amplitude do radar das unidades amostrais visitadas em
campo. Posteriormente esses valores foram transformados para
valores de retroespalhamento. A integra{\c{c}}{\~a}o dos dados
baseou-se na an{\'a}lise da adequa{\c{c}}{\~a}o dos mesmos a
modelos de regress{\~a}o onde a tend{\^e}ncia de comportamento
das vari{\'a}veis (imagem x campo) fosse ajustada. Os resultados
indicaram a exist{\^e}ncia de correla{\c{c}}{\~a}o
significativa entre os dados orbitais e a fitomassa, tendo sido os
modelos de regress{\~a}o mais adequados, aqueles que relacionam
unicamente fitomassa (peso fresco) e coeficiente de
retroespalhamento (imagem do radar), para as {\'a}reas de
savanas, e fitomassa (peso seco) e % SOM (percentagem de sombra),
para as {\'a}reas de florestas. Adicionalmente, com a
utiliza{\c{c}}{\~a}o das informa{\c{c}}{\~o}es espectrais das
imagens-propor{\c{c}}{\~o}es e de t{\'e}cnicas de processamento
digital como segmenta{\c{c}}{\~a}o e classifica{\c{c}}{\~a}o
autom{\'a}tica, foi gerado um mapa tem{\'a}tico dos tipos
fision{\^o}micos da cobertura vegetal da {\'a}rea de estudo,
associado a valores de fitomassa, considerando as
modifica{\c{c}}{\~o}es antr{\'o}picas ocorridas na paisagem.
ABSTRACT: Vegetation coverage quantification is necessary to
provide data for studies related to the processes of global
changes. This study presents biophysical parameters of the
vegetation coverage in the contact region of forests and savannas
in the State of Roraima, Brazil, as it evaluates their
relationships with the orbital data (TM/LANDSAT and SAR/JERS-1).
The study area vegetation communities were described according to
lhe physiognomicstructural and floristic characteristics, and they
are classified in primary forest, secondary succession, savanna
woodland, savanna parkland, and savanna grassland. Through the use
of sensoried information extraction techniques, the products
originating in the optical and microwave ranges were processed and
the digital values were extracted. For the TM sensor image, the
Linear Model of Spectral Mixture technique was applied, generating
the image-fractions (VEG, SOL and SOM), whose digital values were
used in the relationships with lhe biophysical parameters of the
different coverages occurring in the contact area. With
georeferenced JERS and TM images, it was possible to extract
digital values in the radar amplitude image of the sampling units
visited in the field. These digital values were later transformed
in backscattering values. Data integration was based on the
analysis of their adequacy to regression methods where the
variables (image x field) behavior tendency was adjusted. Results
have indicated an existing significant correlation between orbital
data and lhe phytomass, where the most adequate regression models
were those relating exclusively phytomass (fresh weight) and
backscattering values (radar image) for lhe savanna areas, and
phytomass (dry weight) and % SOM (percentage of shade) for the
forest areas. Additionally, with the use of spectral information
on the images-fractions and of digital processing techniques such
as segmentation and automatic classification, a thematic map of
lhe physiognomic types of the study area vegetable coverage was
generated in association with phytomass values, considering the
anthropic changes to the landscape.",
committee = "Santos, Jo{\~a}o Roberto dos (presidente/orientador) and Freitas,
Corina da Costa and Ponzoni, Fl{\'a}vio Jorge and Durigan,
Giselda and Dias, Braulio Ferreira de Souza",
copyholder = "SID/SCD",
englishtitle = "TM/LANDSAT and SAR/JERS data evaluation in characterizing
vegetation coverage and phytomass distribution in the
forest/savanna contact area in the state of Roraima - Brasil",
language = "pt",
pages = "125",
ibi = "8JMKD3MGP3W34T/48BK4K5",
url = "http://urlib.net/ibi/8JMKD3MGP3W34T/48BK4K5",
urlaccessdate = "16 jun. 2024"
}