@InProceedings{SotheGeEsAlScFrLi:2017:AnMuCo,
author = "Sothe, Camile and Gerente, J{\'e}ssica and Escada, Maria Isabel
Sobral and Almeida, Cl{\'a}udia Maria de and Schimalski, Marcos
Benedito and Francisco, Cristiane Nunes and Liesenberg, Veraldo",
affiliation = "{} and {Instituto Nacional de Pesquisas Espaciais (INPE)} and
{Instituto Nacional de Pesquisas Espaciais (INPE)} and {Instituto
Nacional de Pesquisas Espaciais (INPE)}",
title = "An{\'a}lise multitemporal da cobertura vegetal afetada por
movimentos de massa no munic{\'{\i}}pio de Nova Friburgo-RJ",
booktitle = "Anais...",
year = "2017",
editor = "Gherardi, Douglas Francisco Marcolino and Arag{\~a}o, Luiz
Eduardo Oliveira e Cruz de",
pages = "1502--1509",
organization = "Simp{\'o}sio Brasileiro de Sensoriamento Remoto, 18. (SBSR)",
publisher = "Instituto Nacional de Pesquisas Espaciais (INPE)",
address = "S{\~a}o Jos{\'e} dos Campos",
abstract = "In the past, different approaches for automated mass movements
identification based on multispectral orbital images were
developed to focus on the analysis of the spatial distribution of
mass movements occurrences related to distinct triggering events.
However, a continual multi-temporal analysis is important for
monitoring vegetation recovery of affected areas. The first
objective of this paper was to use a semi-automated mapping
approach based on ALOS and RapidEye time series data. For change
detection, a threshold method was applied in a difference image
resulting from the subtraction between NDVI and GNDVI from 2010
and 2011 images. The second objective was to check recovery
vegetation areas incorporating the 2015 image at issue. For this
purpose, NDVI and GNDVI of three images associated with change
objects resulting from the first objective described above were
used in a decision tree classification algorithm. The change
detection approach resulted in the identification of nearly
129-145 ha associated with mass movements occurrence. A
quantitative accuracy assessment for these two methods has
revealed a detection percentage of 75% of mass movements with the
NDVI method, and 67% with the GNDVI method, however, NDVI resulted
in higher commission errors. The classification with C4.5 decision
tree algorithm revealed 121ha of areas under recovery in 2015,
while 106 ha have not been undergone recovery yet. The study
proved the suitability of the developed approaches for efficient
spatiotemporal mass movements mapping areas, representing an
important prerequisite for mass movements hazard and risk
assessment at the regional scale.",
conference-location = "Santos",
conference-year = "28-31 maio 2017",
isbn = "978-85-17-00088-1",
label = "60022",
language = "pt",
organisation = "Instituto Nacional de Pesquisas Espaciais (INPE)",
ibi = "8JMKD3MGP6W34M/3PS4GQN",
url = "http://urlib.net/ibi/8JMKD3MGP6W34M/3PS4GQN",
targetfile = "60022.pdf",
type = "Detec{\c{c}}{\~a}o de mudan{\c{c}}as",
urlaccessdate = "27 abr. 2024"
}