@Article{FurtadoSilvFernNovo:2015:LaCoCl,
author = "Furtado, Luiz Felipe de Almeida and Silva, Thiago Sanna Freire and
Fernandes, Pedro Jos{\'e} Farias and Novo, Evlyn M{\'a}rcia
Le{\~a}o de Moraes",
affiliation = "{Instituto Nacional de Pesquisas Espaciais (INPE)} and
{Universidade Estadual Paulista (UNESP)} and {Universidade Federal
Fluminense (UFF)} and {Instituto Nacional de Pesquisas Espaciais
(INPE)}",
title = "Land cover classification of Lago Grande de Curuai floodplain
(Amazon, Brazil) using multi-sensor and image fusion techniques",
journal = "Acta Amazonica",
year = "2015",
volume = "45",
number = "2",
pages = "195--202",
keywords = "wetlands, remote sensing, synthetic aperture radar, {\'a}reas
{\'u}midas, sensoriamento remoto, radar de abertura
sint{\'e}tica.",
abstract = "Dadas as limita{\c{c}}{\~o}es de diferentes tipos de imagens de
sensores remotos, classifica{\c{c}}{\~o}es autom{\'a}ticas do
uso e cobertura do solo na v{\'a}rzea Amaz{\^o}nica podem
resultar em {\'{\i}}ndices de acur{\'a}cia
insatisfat{\'o}rios. Uma das maneiras de melhorar esses
{\'{\i}}ndices {\'e} atrav{\'e}s da combina{\c{c}}{\~a}o de
dados de distintos sensores, por fus{\~a}o de imagens ou
atrav{\'e}s de classifica{\c{c}}{\~o}es multi-sensores. Desta
forma, o presente estudo teve o objetivo de determinar qual
m{\'e}todo de classifica{\c{c}}{\~a}o {\'e} mais eficiente em
melhorar os {\'{\i}}ndices de acur{\'a}cia das
classifica{\c{c}}{\~o}es do uso e cobertura do solo para a
v{\'a}rzea Amaz{\^o}nica e {\'a}reas {\'u}midas similares -
(a) a fus{\~a}o sint{\'e}tica de imagens SAR e {\'o}pticas ou
(b) a classifica{\c{c}}{\~a}o multi-sensor de imagens
{\'o}pticas e SAR pareadas. Classifica{\c{c}}{\~o}es da
cobertura do solo com base em imagens de um {\'u}nico sensor
(Landsat TM ou Radarsat-2) foram comparadas com as
classifica{\c{c}}{\~o}es multi-sensor e
classifica{\c{c}}{\~o}es baseadas em fus{\~a}o de imagens. A
an{\'a}lise de imagens baseada em objetos (OBIA) e o algoritmo de
minera{\c{c}}{\~a}o de dados J.48 foram utilizados para realizar
a classifica{\c{c}}{\~a}o autom{\'a}tica, cuja precis{\~a}o
foi avaliada com o {\'{\i}}ndice kappa e com as medidas de
discord{\^a}ncia de aloca{\c{c}}{\~a}o e de quantidade,
recentemente propostas na literatura. Em geral, as
classifica{\c{c}}{\~o}es baseadas em imagens {\'o}pticas
apresentaram melhor precis{\~a}o do que as
classifica{\c{c}}{\~o}es baseadas em dados SAR. Uma vez que
ambos os conjuntos de dados foram combinados em uma abordagem
multi-sensores, houve uma redu{\c{c}}{\~a}o de 2% no erro de
aloca{\c{c}}{\~a}o da classifica{\c{c}}{\~a}o, uma vez que o
m{\'e}todo foi capaz de superar parte das limita{\c{c}}{\~o}es
presentes em ambas as imagens. Contudo, a precis{\~a}o diminuiu
quando foram usados m{\'e}todos de fus{\~a}o de imagens.
Concluiu-se que o m{\'e}todo de classifica{\c{c}}{\~a}o
multi-sensor {\'e} mais apropriado para
classifica{\c{c}}{\~o}es de uso do solo na v{\'a}rzea
amaz{\^o}nica. ABSTRACT: Given the limitations of different types
of remote sensing images, automated land-cover classifications of
the Amazon v{\'a}rzea may yield poor accuracy indexes. One way to
improve accuracy is through the combination of images from
different sensors, by either image fusion or multi-sensor
classifications. Therefore, the objective of this study was to
determine which classification method is more efficient in
improving land cover classification accuracies for the Amazon
v{\'a}rzea and similar wetland environments - (a) synthetically
fused optical and SAR images or (b) multi-sensor classification of
paired SAR and optical images. Land cover classifications based on
images from a single sensor (Landsat TM or Radarsat-2) are
compared with multi-sensor and image fusion classifications.
Object-based image analyses (OBIA) and the J.48 data-mining
algorithm were used for automated classification, and
classification accuracies were assessed using the kappa index of
agreement and the recently proposed allocation and quantity
disagreement measures. Overall, optical-based classifications had
better accuracy than SAR-based classifications. Once both datasets
were combined using the multi-sensor approach, there was a 2%
decrease in allocation disagreement, as the method was able to
overcome part of the limitations present in both images. Accuracy
decreased when image fusion methods were used, however. We
therefore concluded that the multi-sensor classification method is
more appropriate for classifying land cover in the Amazon
v{\'a}rzea.",
doi = "10.1590/1809-4392201401439",
url = "http://dx.doi.org/10.1590/1809-4392201401439",
issn = "0044-5967",
language = "en",
targetfile = "furtado_land cover.pdf",
urlaccessdate = "15 jun. 2024"
}