@Article{NevesKöFoSoGiHe:2021:HiMaBr,
author = "Neves, Alana K. and K{\"o}rting, Thales Sehn and Fonseca, Leila
Maria Garcia and Soares, Anderson Reis and Girolamo Neto, Cesare
Di and Heipke, Christian",
affiliation = "{Instituto Nacional de Pesquisas Espaciais (INPE)} and {Instituto
Nacional de Pesquisas Espaciais (INPE)} and {Instituto Nacional de
Pesquisas Espaciais (INPE)} and {Instituto Nacional de Pesquisas
Espaciais (INPE)} and {Instituto Nacional de Pesquisas Espaciais
(INPE)} and {Leibniz Universit{\"a}t Hannover}",
title = "Hierarchical mapping of Brazilian Savanna (Cerrado) physiognomies
based on deep learning",
journal = "Journal of Applied Remote Sensing",
year = "2021",
volume = "15",
number = "04",
pages = "e044504",
keywords = "Savanna, Cerrado, physiognomy, semantic segmentation, spectral
channels, protected area.",
abstract = "The Brazilian Savanna, also known as Cerrado, is considered a
global hotspot for biodiversity conservation. The detailed mapping
of vegetation types, called physiognomies, is still a challenge
due to their high spectral similarity and spatial variability.
There are three major ecosystem groups (forest, savanna, and
grassland), which can be hierarchically subdivided into 25
detailed physiognomies, according to a well-known classification
system. We used an adapted U-net architecture to process a
WorldView-2 image with 2-m spatial resolution to hierarchically
classify the physiognomies of a Cerrado protected area based on
deep learning techniques. Several spectral channels were tested as
input datasets to classify the three major ecosystem groups (first
level of classification). The dataset composed of RGB bands plus
2-band enhanced vegetation index (EVI2) achieved the best
performance and was used to perform the hierarchical
classification. In the first level of classification, the overall
accuracy was 92.8%. On the other hand, for the savanna and
grassland detailed physiognomies (second level of classification),
86.1% and 85.0% were reached, respectively. As the first work that
intended to classify Cerrado physiognomies in this level of detail
using deep learning, our accuracy rates outperformed others that
applied traditional machine learning algorithms for this task. ©
The Authors. Published by SPIE under a Creative Commons
Attribution 4.0 International License. Distribution or
reproduction of this work in whole or in part requires full
attribution of the original publication, including its DOI. [DOI:
10.1117/1.JRS.15.044504].",
doi = "10.1117/1.JRS.15.044504",
url = "http://dx.doi.org/10.1117/1.JRS.15.044504",
issn = "1931-3195",
label = "lattes: 5123287769635741 3 NevesK{\"o}FoSoGiHe:2021:HiMaBr",
language = "pt",
targetfile = "neves_hierarchical.pdf",
url = "https://caps.luminad.com:8443/stockage/stock/SPIE/LDL-SPIE-JARS-210442/JARS-210442_online.pdf",
urlaccessdate = "21 maio 2024"
}