@PhDThesis{Ferri:2002:UtReEs,
author = "Ferri, Clotilde Pinheiro",
title = "Utiliza{\c{c}}{\~a}o da reflect{\^a}ncia espectral para a
estimativa de pigmentos fotossint{\'e}ticos em doss{\'e}is de
soja [Gycine Max (.L), Merril]",
school = "Instituto Nacional de Pesquisas Espaciais (INPE)",
year = "2002",
address = "S{\~a}o Jos{\'e} dos Campos",
month = "2002-03-27",
keywords = "agronomia, pigmentos, fotoss{\'{\i}}ntese, reflect{\^a}ncia
espectral, doss{\'e}is, reflect{\^a}ncia, resolu{\c{c}}{\~a}o
espectral, clorofila, soja, pigments, photosynthesis, spectral
reflectance, canopies (vegetation), reflectance, spectral
resolution, chlorophylls, soybeans.",
abstract = "Os pigmentos fotossint{\'e}ticos s{\~a}o essenciais para o
desenvolvimento das plantas, pois s{\~a}o respons{\'a}veis pela
captura da energia solar incidente usada na fotoss{\'{\i}}ntese.
Com o desenvolvimento do sensoriamento remoto hiperespectral,
tem-se aberto a possibilidade de quantificar estes pigmentos
individualmente em grandes extens{\~o}es de lavouras
agr{\'{\i}}colas e estas informa{\c{c}}{\~o}es auxiliam na
determina{\c{c}}{\~a}o do estado fisiol{\'o}gico da
vegeta{\c{c}}{\~a}o, na discrimina{\c{c}}{\~a}o de
esp{\'e}cies e na estimativa da produtividade. Uma das maneiras
de utilizar os dados de sensores hiperespectrais na
determina{\c{c}}{\~a}o da concentra{\c{c}}{\~a}o de pigmentos,
{\'e} pelo uso de {\'{\i}}ndices de reflect{\^a}ncia espectral
que utilizam bandas espectrais estreitas na forma de soma,
raz{\~a}o ou multiplica{\c{c}}{\~a}o. Assim, com este trabalho,
buscou-se fornecer bases para a constru{\c{c}}{\~a}o de
conhecimentos na {\'a}rea de sensoriamento remoto hiperespectral
e suas rela{\c{c}}{\~o}es com culturas agr{\'{\i}}colas. Como
hip{\'o}tese de trabalho, prop{\~o}e-se que {\'e}
poss{\'{\i}}vel, usando sensoriamento remoto de alta
resolu{\c{c}}{\~a}o espectral, estimar pigmentos
fotossint{\'e}ticos de doss{\'e}is vegetais agr{\'{\i}}colas,
sendo necess{\'a}rio levar em conta as influ{\^e}ncias de solos
espectralmente distintos. A presente pesquisa sugere tamb{\'e}m
que existem rela{\c{c}}{\~o}es significativas entre
reflect{\^a}ncia espectral e par{\^a}metros agron{\^o}micos de
uma cultura agr{\'{\i}}cola, tendo como objetivos: (a) avaliar
as rela{\c{c}}{\~o}es entre componentes vegetais da cultura da
soja e resposta espectral atrav{\'e}s dos {\'{\i}}ndices
espectrais: R750/R700, R750/R550, Ratio Analysis of Reflectance
Spectra (RARS), Pigment Specific Simple Ratio (PSSR)e Pigment
Specific Normalized Difference (PSND); e (b) analisar a
influ{\^e}ncia de dois solos espectralmente diferentes no
comportamento espectral dessa cultura. Para testar a hip{\'o}tese
estabelecida e atingir os objetivos propostos, a presente pesquisa
foi realizada por m{\'e}todo experimental, em
condi{\c{c}}{\~o}es de casa de vegeta{\c{c}}{\~a}o,
utilizando-se espectroradi{\^o}metros de alta
resolu{\c{c}}{\~a}o, com a cultura da soja [Glycine max (L.),
Merril] tendo sido conduzida e monitorada espectralmente ao longo
do seu ciclo fenol{\'o}gico. Assim, em fun{\c{c}}{\~a}o dos
procedimentos utilizados e das an{\'a}lises realizadas, foi
poss{\'{\i}}vel chegar a algumas conclus{\~o}es e
recomenda{\c{c}}{\~o}es, verificando que (a) as plantas de soja
tiveram um desenvolvimento considerado normal mesmo tendo sido
cultivadas em condi{\c{c}}{\~o}es de casa de
vegeta{\c{c}}{\~a}o, tendo apresentado um sistema
assimilat{\'o}rio suficiente e eficiente para a
produ{\c{c}}{\~a}o e ac{\'u}mulo de mat{\'e}ria seca; (b) os
solos usados, mesmo tendo sido dois extremos espectrais (o
primeiro com caracter{\'{\i}}sticas de baixa reflect{\^a}ncia e
o segundo com alta reflect{\^a}ncia n{\~a}o influenciaram nas
respostas espectrais dos doss{\'e}is de soja na maior parte do
seu ciclo fenol{\'o}gico; (c) a partir da fase fenol{\'o}gica
V3, da cultura as respostas espectrais j{\'a} come{\c{c}}aram a
ser de vegeta{\c{c}}{\~a}o, tendo iniciado ent{\~a}o a
manifesta{\c{c}}{\~a}o de rela{\c{c}}{\~o}es entre as
respostas espectrais e a concentra{\c{c}}{\~a}o de pigmentos,
confirmando a validade de se estudar estas rela{\c{c}}{\~o}es
para posteriores determina{\c{c}}{\~o}es do est{\'a}dio
fenol{\'o}gico e de outras caracter{\'{\i}}sticas de vigor,
sanidade e produtividade da cultura por imagens hiperespectrais;
(d) todos os espectros coletados apresentaram maior
reflect{\^a}ncia no infravermelho pr{\'o}ximo, em torno de 740
nm; (e) a posi{\c{c}}{\~a}o da borda vermelha situou-se em torno
de 700 nm; (f) os espectros apresentaram curvas
caracter{\'{\i}}sticas de acordo com a fase de desenvolvimento
da cultura da soja, tendo influ{\^e}ncia do solo de fundo nas
fases iniciais e finais da cultura; (g) as posi{\c{c}}{\~o}es do
comprimento de onda equivalente ao valor m{\'{\i}}nimo de
reflect{\^a}ncia no vermelho (Vmin) e do ponto de inflex{\~a}o
(PI) da curva apresentaram boas rela{\c{c}}{\~o}es com o
conte{\'u}do de pigmentos, tendo o Vmin diminu{\'{\i}}do com o
aumento da concentra{\c{c}}{\~a}o de Clorofila a e o PI
aumentado com o aumento da concentra{\c{c}}{\~a}o de Clorofila
total; (h) as raz{\~o}es R750/R550 e R750/R700 mostraram ser bons
{\'{\i}}ndices para a determina{\c{c}}{\~a}o de Clorofila a e
de Clorofila total, e estes {\'{\i}}ndices, quando comparados
com o NDVI (um dos {\'{\i}}ndices espectrais de banda larga mais
usados em sensoriamento remoto da vegeta{\c{c}}{\~a}o),
mostraram ser mais eficientes; (i) a an{\'a}lise dos tr{\^e}s
{\'{\i}}ndices de bandas estreitas, o RARS, o PSND e o PSSR,
para a determina{\c{c}}{\~a}o de Clorofila a e b e
Caroten{\'o}ides, mostraram que o melhor {\'{\i}}ndice foi o
PSSR; (j) o {\'{\i}}ndice RARS n{\~a}o mostrou boas
rela{\c{c}}{\~o}es quando os dados de todo o ciclo
fenol{\'o}gico foram usados; por{\'e}m, apresentou bons
resultados quando se avaliou a cultura em fases bem delimitadas,
como a vegetativa e a reprodutiva separadamente, para a
determina{\c{c}}{\~a}o de Clorofila a e b. Para a
determina{\c{c}}{\~a}o de Caroten{\'o}ides, este
{\'{\i}}ndice n{\~a}o foi eficiente; (k) o {\'{\i}}ndice PSND
tamb{\'e}m apresentou ter rela{\c{c}}{\~a}o com o conte{\'u}do
de Clorofila a e b e nenhuma rela{\c{c}}{\~a}o com o
conte{\'u}do de Caroten{\'o}ides; (l) somente o {\'{\i}}ndice
PSSR foi um bom estimador da concentra{\c{c}}{\~a}o de pigmentos
nas plantas, em n{\'{\i}}vel de dossel, sendo que este
{\'{\i}}ndice tamb{\'e}m n{\~a}o apresentou bons resultados
para estimar a concentra{\c{c}}{\~a}o de Caroten{\'o}ides.
Assim, com o presente trabalho fica comprovado que {\'e}
poss{\'{\i}}vel estimar pigmentos fotossint{\'e}ticos
atrav{\'e}s de Sensoriamento Remoto Hiperespectral em
n{\'{\i}}vel de dossel, atrav{\'e}s dos {\'{\i}}ndices
avaliados e de outros que podem ser gerados. Trabalhos devem ser
realizados com o prop{\'o}sito de se avaliar estes e outros
{\'{\i}}ndices em n{\'{\i}}vel de dossel, tanto em
laborat{\'o}rio como em condi{\c{c}}{\~o}es de campo, usando
espectroradi{\^o}metros e imagens hiperespectrais a fim de se
verificar a efici{\^e}ncia destes para futuras
utiliza{\c{c}}{\~o}es agr{\'{\i}}colas. ABSTRACT:
Photosynthetic pigments are essential for plant development,
because they are responsible for the incident solar energy
absorption used in photosynthesis. Individually quantifying these
pigments in great extensions of agricultural crops has been made
possible by hyperspectral remote sensing development and these
information can be used to produce a more accurate vegetation
physiologic state determination, species discrimination and
productivity estimation. Pigments concentration can be determined
through the use of reflectance indices which utilize narrow
spectral bands in sum, reason or multiplication forms. The present
work intends to establish the basis in hyperspectral remote
sensing area and its relationships with agricultural cultures,
supposing that with the usage of high spectral resolution remote
sensing and taking into account the spectrally different soil
influences it is possible to estimate photosynthetic pigment
content in agricultural vegetable canopies. The present work also
suggests the existence of important relations between spectral
reflectance and the agricultural culture agronomic parameters with
the following main objectives: (a) evaluation of spectral response
by means of spectral indices: R750/R700, R750/R550, Ratio Analysis
of Reflectance Spectra (RARS), Pigment Specific Simple Ratio
(PSSR) and Pigment Specific Normalized Difference (PSND) and
soybean culture vegetable components relationships; and (b)
analysis of two spectrally different soils influence in soybean
culture spectral behavior. Experimental methods, in vegetation
house conditions, were used to verify basic assumptions and along
its phenologic cycle the soybean culture [Glycine max (L.),
Merril] was monitored with high resolution spectroradiometer. So
considering the methods and the performed analysis it is possible
to state some conclusions and recomendations taking into account
that (a) although cultivated in vegetation house conditions the
soybean plants grew as expected presenting dry matter accumulation
and production due to an enough and efficient assimilatory system;
(b) during the soybean canopy phenological cycle no soil influence
was observed in the spectral response even using two specttral
kinds of soils: the first one with low relfectance characteristic
and the second one with high reflectance characteristic; (c)
typical vegetation spectral response and relationships between
these responses and pigment concentration started from V3
phenologic phase confirming these relations worth for further
determinations of phenologic stage and characteristics like vigor,
sanity and productivity from spectral images; (d) the greater
reflectance of all collected spectra was found to be in the near
infrared region around 740nm; (e) the red edge position was
determined around 700 nm; (f) spectra characteristic curves
changed according to soybean culture development phase and at
initial and final phases a background soil influence was observed;
(g) a significant relation between pigment content, the minimum
red reflectance (Vmin) and the inflexion point (PI) was observed
and it was also noted that the greater the Chlorophyll a
concentration the lower the Vmin and the greater the PI the lower
the total chlorophyll; (h) R750/R550 and R750/R700 ratios are
important indices for Chlorophyll a and total chlorophyll
determination and very effective when compared with NDVI (one of
the vegetation remote sensing most used wide band indices); (i)
RARS, PSND and PSSR narrow band indices analysis showed that PSSR
is the better one for Chlorophyll a and b and Carotenoids
determination; (j) when using the whole phenological cycle data
the RARS indice is not efficient although it leads to good results
when used in small parts of the cycle like the vegetative or the
reproductive ones. This indice is also not efficient for
Chlorophyll a and b and Carotonoids determination; (k) there is a
relationship between PSND indice and chlorophyll a and b and no
relationship between this indice and Carotenoids content; (l) At
canopy level a reliable plant pigment content prediction can only
be achieved from PSSR indice although this indice is not efficient
for Carotenoids content determination. So with this work it can be
stated that photosynthetic pigment content measurements at canopy
level can be performed by hyperspectral remote sensing in a
reliable way through evaluated indices and others that can be
generated and further studies should be carried out with these and
other indices with evaluating purposes at canopy level either in
laboratory as in field conditions using spectroradio.",
committee = "Ponzoni, Fl{\'a}vio Jorge (presidente) and Formaggio,
Ant{\^o}nio Roberto (orientador) and Galv{\~a}o, L{\^e}nio
Soares and Schiavinato, Marlene Aparecida and Souza Filho, Carlos
Roberto de and Dematt{\^e}, Jos{\'e} Alexandre Melo",
englishtitle = "Soybean canopy [Glycine max (L.), Merril] photosynthetic pigments
estimation by measns of spectral reflectance",
label = "9944",
language = "pt",
pages = "173",
ibi = "83LX3pFwXQZ5Jpy/wMBwH",
url = "http://urlib.net/ibi/83LX3pFwXQZ5Jpy/wMBwH",
targetfile = "publicacao.pdf",
urlaccessdate = "05 maio 2024"
}