@MastersThesis{Dória:2015:CaDiFl,
author = "D{\'o}ria, Vinicius Etchebeur Medeiros",
title = "Sensoriamento remoto de luzes noturnas para estimativas
populacionais em escalas regional e local: os casos do Distrito
Florestal Sustent{\'a}vel da BR-163 (PA) e da Regi{\~a}o
Metropolitana de S{\~a}o Paulo",
school = "Instituto Nacional de Pesquisas Espaciais (INPE)",
year = "2015",
address = "S{\~a}o Jos{\'e} dos Campos",
month = "2015-11-30",
keywords = "luzes noturnas, OLS, VIIRS, estimativa de popula{\c{c}}{\~a}o,
an{\'a}lise espacial, nighttime light, population estimates,
spatial analysis.",
abstract = "O processo de urbaniza{\c{c}}{\~a}o na Amaz{\^o}nia
contempor{\^a}nea se expressa n{\~a}o somente nas cidades e
vilas estabelecidas e estruturadas, mas tamb{\'e}m em outras
formas de organiza{\c{c}}{\~a}o do espa{\c{c}}o decorrentes de
atividades humanas. Por sua vez, Regi{\~o}es Metropolitanas
experimentam variados processos locais de expans{\~a}o urbana,
com ocupa{\c{c}}{\~o}es que ocorrem horizontal e verticalmente.
Dados de sensoriamento remoto e t{\'e}cnicas de geoprocessamento
t{\^e}m sido utilizados para descrever e entender processos de
urbaniza{\c{c}}{\~a}o e ocupa{\c{c}}{\~a}o do territ{\'o}rio
e, mais recentemente, para auxiliar na caracteriza{\c{c}}{\~a}o
da distribui{\c{c}}{\~a}o espacial da popula{\c{c}}{\~a}o
nestes territ{\'o}rios urbanizados e na estimativa do volume
populacional. Este trabalho tem por objetivo avaliar a
contribui{\c{c}}{\~a}o de imagens mosaico de luzes noturnas do
\emph{Operational Linescan System} (OLS) e do \emph{Visible
Infrared Imaging Radiometer Suite} (VIIRS) para estimativas de
popula{\c{c}}{\~a}o, estudando as {\'a}reas do Distrito
Florestal Sustent{\'a}vel da BR163 (DFS-BR163) na Amaz{\^o}nia
Paraense, e a Regi{\~a}o Metropolitana de S{\~a}o Paulo (RMSP).
Dados de luzes noturnas e dados demogr{\'a}ficos foram
sistematizados para inferir o quantitativo populacional e a
distribui{\c{c}}{\~a}o da popula{\c{c}}{\~a}o em escalas
regional e local. Realizou-se inicialmente a
intercalibra{\c{c}}{\~a}o das imagens mosaico de luzes do OLS e
o refinamento da imagem mosaico de luzes do VIIRS. Para verificar
a utilidade da s{\'e}rie temporal dos mosaicos de luzes noturnas
em estimativas de popula{\c{c}}{\~a}o, a t{\'e}cnica de
regress{\~a}o linear foi adotada para o DFS-BR163, na
an{\'a}lise de imagens mosaico OLS de 2000, 2007 e 2010. A
Regress{\~a}o Geograficamente Ponderada (GWR) foi aplicada para
imagem mosaico VIIRS 2012 em escala local e intraurbana na RMSP.
Da an{\'a}lise dos resultados obtidos para a escala do DFS-BR163
pode-se afirmar que: (1) a intercalibra{\c{c}}{\~a}o das imagens
mosaico do OLS deve ser adaptada para as condi{\c{c}}{\~o}es
locais para garantir an{\'a}lises consistentes da s{\'e}rie
temporal; (2) o refinamento da imagem mosaico VIIRS {\'e}
essencial para eliminar as luzes n{\~a}o associadas {\`a}
presen{\c{c}}a de ocupa{\c{c}}{\~a}o humana; (3) a capacidade
do sensor VIIRS em identificar e individualizar {\'a}reas de
diferentes intensidades de ocupa{\c{c}}{\~a}o humana indica o
potencial destas imagens para estudos de distribui{\c{c}}{\~a}o
espacial da popula{\c{c}}{\~a}o; (4) os dados indicaram
regress{\~o}es com fun{\c{c}}{\~o}es lineares, para os mosaicos
do OLS, e exponencial, para o mosaico do VIIRS, sendo que de 67\%
a 80\% da variabilidade observada nos quantitativos populacionais
p{\^o}de ser explicada pelas imagens de luzes noturnas, os erros
das estimativas s{\~a}o explicados por fatores locais das
{\'a}reas iluminadas e as estimativas do VIIRS apresentaram erros
significativamente menores do que as do OLS. Para a RMSP,
constatou-se que a GWR apresentou estimativas populacionais
satisfat{\'o}rias, com erro geral de 5\% em rela{\c{c}}{\~a}o
{\`a} popula{\c{c}}{\~a}o observada, e os coeficientes da GWR
mostraram-se associados a diferentes padr{\~o}es de
ocupa{\c{c}}{\~a}o urbana na RMSP. Por fim, este trabalho indica
o potencial do uso de dados de luzes noturnas do OLS e do VIIRS
para estimativas de popula{\c{c}}{\~a}o nas escalas regional e
local, podendo auxiliar os modelos de distribui{\c{c}}{\~a}o e
de estimativa anual da popula{\c{c}}{\~a}o. ABSTRACT: The
contemporary urbanization process in the Amazon region is
expressed not only by structured cities and established villages,
but also by other spatial forms of territory occupation generated
by human activities. On the other hand, Metropolitan Regions
experience various local urban growing processes that occur
horizontally and vertically. Remote sensing data and
geoinformation techniques have been used to describe and
understand urbanization and colonization processes, and,
particularly in recent years, they have been used to assist on the
characterization of spatial population distribution in these
urbanized areas and on population estimates. This work aims to
identify the contribution of nighttime lights images from
Operational Linescan System (OLS) and Visible Infrared Imaging
Radiometer Suite (VIIRS) for population estimates in the
BR163-Sustainable Forest District (DFS-BR163), in the Brazilian
Amazon, and the S{\~a}o Paulo Metropolitan Region (RMSP).
Nighttime lights images and demographic data were organized to
infer population counts and distribution, considering both
regional and local scales. First, intercalibraton of OLS nighttime
lights mosaic images and refinement of VIIRS nighttime lights
mosaic image to remove noise and lights from ephemeral events were
carried out. Then, a linear regression model explored the
relationships between OLS nighttime lights mosaic images and
population estimates in DFS-BR163, in 2000, 2007 and 2010. At
local and intraurban scale, Geographically Weighted Regression
(GWR) was used to study VIIRS 2012 mosaic image and population
distribution in the RMSP. The results from DFS-BR163 scale suggest
that: (1) the OLS mosaic images intercalibration has to be adapted
to the local conditions to guarantee consistent analysis of the
time series; (2) the refinement of VIIRS nighttime lights mosaic
image is essential to eliminate ephemeral sources and noises; (3)
the VIIRS sensibility to identify and individualize human
settlements of different intensities of human occupation
highlights the potential of VIIRS images for spatial population
distribution studies; (4) population estimates presented linear
relationship for OLS mosaic images regressions and exponential
relationship for VIIRS. Also, 67\% to 80\% of the population
variability could be explained by nighttime lights images; the
estimate errors were related to local characteristics of lit
areas; and VIIRS estimate errors were substantially lower than OLS
s. For the RMSP, GWR presented satisfactory estimates, with
general error of 5\% for the observed population data, and the
GWR coefficients were related to different urban occupation
patterns in the RMSP. Lastly, this work indicates the potential of
OLS and VIIRS nighttime lights data for population estimates at
both regional and local scales, suggesting that these images could
also be used as ancillary data for spatial distribution models and
annual population estimates.",
committee = "Monteiro, Ant{\^o}nio Miguel Vieira (presidente/orientador) and
Kampel, Silvana Amaral (orientadora) and Almeida, Cl{\'a}udia
Maria de and Novo, Evlyn M{\'a}rcia Le{\~a}o de Moraes and
Bueno, Maria do Carmo Dias",
copyholder = "SID/SCD",
englishtitle = "Nighttime light images for population estimates at regional and
local scales: the cases of the Sustainable Forest District of
BR-163 (PA) and the S{\~a}o Paulo Metropolitan Region",
language = "pt",
pages = "131",
ibi = "8JMKD3MGP3W34P/3KLCMC5",
url = "http://urlib.net/ibi/8JMKD3MGP3W34P/3KLCMC5",
targetfile = "publicacao.pdf",
urlaccessdate = "27 abr. 2024"
}