@MastersThesis{Macedo:2009:LaAe,
author = "Macedo, Rodrigo de Campos",
title = "Estimativa volum{\'e}trica de povoamento clonal de Eucalyptus sp.
atrav{\'e}s de laserscanner aerotransportado",
school = "Instituto Nacional de Pesquisas Espaciais (INPE)",
year = "2009",
address = "S{\~a}o Jos{\'e} dos Campos",
month = "2009-02-27",
keywords = "Invent{\'a}rio florestalk, LiDAR, laserscanner, modelo digital de
eleva{\c{c}}{\~a}o, estrutura florestal, forest inventor, LiDAR,
laserscanner, digital elevation model, forest structure.",
abstract = "O gerenciamento florestal requer dados dendrom{\'e}tricos e
topogr{\'a}ficos como subs{\'{\i}}dio aos processos de tomada
de decis{\~a}o. Os meios mais utilizados para isto s{\~a}o os
invent{\'a}rios e levantamentos de campo. Com o aprimoramento dos
SIGīs (Sistemas de Informa{\c{c}}{\~o}es Geogr{\'a}ficas) e o
aumento das op{\c{c}}{\~o}es de sensoriamento remoto, tais como
sensores ativos e passivos de alta resolu{\c{c}}{\~a}o espacial,
a demanda pela aquisi{\c{c}}{\~a}o e pelo processamento de dados
dendrom{\'e}tricos e topogr{\'a}ficos, se acentuou. Em meados de
1993, sensores opticamente ativos, conhecidos como laserscanners
come{\c{c}}aram a ser utilizados especificamente no meio
florestal, visando modelar o terreno e estimar vari{\'a}veis
dendrom{\'e}tricas, como a altura dos povoamentos e, mais
recentemente, a quantidade de indiv{\'{\i}}duos arb{\'o}reos, a
{\'a}rea de copa e o volume madeireiro, principalmente
atrav{\'e}s da segmenta{\c{c}}{\~a}o do dossel florestal. A
hip{\'o}tese {\'e} que a estimativa volum{\'e}trica realizada
com dados obtidos via laserscanner aerotransportado {\'e}
compat{\'{\i}}vel com a estimativa proveniente do
invent{\'a}rio florestal tradicional. O objetivo principal {\'e}
avaliar o potencial da utiliza{\c{c}}{\~a}o de laserscanner
aerotransportado para estimativa volum{\'e}trica de povoamentos
clonais de Eucalyptus sp., em fazendas localizadas em relevo
ondulado. Foram sobrevoados 1.554,7 ha em 6 fazendas utilizadas
para plantio de Eucaliptos. Dentre as fazendas sobrevoadas,
elegeu-se uma F154-Rosa Helena para registrar neste documento e
realizar as valida{\c{c}}{\~o}es necess{\'a}rias. Localiza-se
no munic{\'{\i}}pio de Igarat{\'a}/SP e possui 145,46ha de
plantio efetivo com 4 anos de idade. Como sensor, foi utilizado o
Optech ALTM 2050, com v{\^o}o de 1.000m de altura, footprint de
0,25m. A precis{\~a}o posicional final de 0,12m nas coordenadas X
e Y e 0,21m em Z, compat{\'{\i}}veis com as
caracter{\'{\i}}sticas dos produtos validados. A
diferen{\c{c}}a entre as cotas registradas no Modelo Digital de
Terreno e mensuradas em campo foi de 0,44m. O erro
altim{\'e}trico nominal de um aerolevantamento LiDAR de 1.000m de
altura {\'e} de 0,15m para locais sem cobertura vegetal.
Considerando que n{\~a}o houve presen{\c{c}}a de
vegeta{\c{c}}{\~a}o herb{\'a}cea e n{\~a}o houve
diferencia{\c{c}}{\~a}o entre os RMSīs obtidos nos perfis com e
sem cobertura, em n{\'{\i}}vel ou em desn{\'{\i}}vel, o
detalhamento presente no micro-relevo ocasiona o erro posicional
planim{\'e}trico, que por sua vez, influencia na medida
altim{\'e}trica. De acordo com as compara{\c{c}}{\~o}es
realizadas, os modelos gerados podem ser utilizados para a
obten{\c{c}}{\~a}o de vari{\'a}veis dendrom{\'e}tricas,
requisitos para a modelagem volum{\'e}trica. Implementou-se um
algoritmo de extra{\c{c}}{\~a}o de vari{\'a}veis
dendrom{\'e}tricas e, considerando todas as parcelas medidas em
campo, atingiu-se um percentual de erro de -3,52% para quantidade
de {\'a}rvores, -2,26% para altura, 19,36% para {\'a}rea de
copa, 5,33% para DAP e -1,8% para volume. A partir da an{\'a}lise
dos dados foi poss{\'{\i}}vel verificar que apesar do algoritmo
subestimar a quantidade de {\'a}rvores e a altura, e superestimar
a {\'a}rea de copa e o DAP, a estimativa volum{\'e}trica
manteve-se dentro de um intervalo aceit{\'a}vel, tanto em termos
de intervalo de confian{\c{c}}a, quanto em erro amostral. Diante
dos resultados obtidos, aceita-se a hip{\'o}tese inicial, ou
seja, a estimativa volum{\'e}trica realizada com dados obtidos
via laserscanner aerotransportado {\'e} compat{\'{\i}}vel com a
estimativa proveniente do invent{\'a}rio florestal tradicional.
ABSTRACT: Forest management activities usually rely on
dendrometric and topographic data as a means to support their
decision making processes. For this end, forest inventory and
topographic survey are commonly used. The recent advances in GIS
and the improvements in the spatial resolution of remote sensing
systems, both passive and active ones, largely increased the
demand for the acquisition and processing of dendrometric and
topographic data. Since 1993, optical active sensors \−
known as laserscanners \− started to be specifically used
in forest applications for ground modeling and dendrometric
parameters estimates, e.g. tree height, and more recently,
quantity of trees, crown area, and wood volume, mainly in the
segmentation process of the forest canopy. The core hypothesis of
this research is that the volumetric estimate carried out with
LiDAR data is compatible with estimates of traditional forest
inventory techniques (based on ground data). The main objective is
then to evaluate the potential of using airborne laserscanner for
volumetric estimate of clonal plantation of Eucalyptus sp., found
in steep slope regions. Six farms were select for this study, of
which a total area of 1,554.7 ha was sampled. One farm F154-Rosa
Helena was chosen for detailed records and validation. Such farm
is located in the municipality of Igarat{\'a} (Sao Paulo State)
and owns a four-year old plantation, distributed over an effective
area of 145.46 ha. The Optech ALTM 2050 has been employed, with a
0.25 m footprint, acquired by an average flight height of 1,000 m.
Its planimetric positional accuracy was 0.12 m (X, Y) and the
elevation positional accuracy was 0.21 m (Z). The difference
between the elevations of the Digital Terrain Model and the ones
obtained by field measurements was of 0.44 m. The nominal
elevation error of an airborne LiDAR (with a flight height of
1,000m) is 0.15 m in the case of areas with no vegetation cover.
Considering that the study area did not present herbaceous
vegetation and that no meaningful difference was observed between
the root mean square errors of profiles with and without
vegetation cover (both on flat and uneven terrains), it can be
stated that the details found in the micro-relief of the study
area are to a large extent responsible for the impact of the
planimetric errors in the elevation measurements. In general, the
maximum height values of the Digital Height Model were smaller
than those of the ground data. The generated models can be used
for the extraction of dendrometric parameters, necessary for
volumetric modeling. A special algorithm for the extraction of
dendrometric parameters was implemented using the IDL environment.
Considering all plots regarded in this analysis, the error share
was 3.52% for the total number of trees; 2.26% for tree height;
19.36% for crown area; 5.33% for Diameter Breast Height (DBH), and
98.2% for volume. The algorithm underestimated the number of trees
and their height and overestimated the crown area and DBH. The
volumetric estimate is though acceptable. The core hypothesis of
this research is thus accepted; since the volumetric estimate
executed with airborne laserscanner data is compatible with
estimates derived from the use of traditional forest inventory
techniques.",
committee = "Soares, Jo{\~a}o Vianei (presidente/orientador) and Santos,
Jo{\~a}o Roberto dos (orientador) and Ponzoni, Fl{\'a}vio Jorge
and Mitishita, Edson Aparecido",
copyholder = "SID/SCD",
englishtitle = "Volumetric estimate of Eucalyptus sp. clonal plantation using
airborne laserscanner",
language = "pt",
pages = "143",
ibi = "8JMKD3MGP8W/34NUA58",
url = "http://urlib.net/ibi/8JMKD3MGP8W/34NUA58",
targetfile = "publicacao.pdf",
urlaccessdate = "03 maio 2024"
}