@InProceedings{EberhardtLuFoSaScTr:2015:DeÁrAg,
author = "Eberhardt, Isaque Daniel Rocha and Luiz, Alfredo Jos{\'e} Barreto
and Formaggio, Ant{\^o}nio Roberto and Sanches, Ieda Del Arco and
Schultz, Bruno and Trabaquini, Kleber",
affiliation = "{} and {} and {Instituto Nacional de Pesquisas Espaciais (INPE)}
and {Instituto Nacional de Pesquisas Espaciais (INPE)} and
{Instituto Nacional de Pesquisas Espaciais (INPE)}",
title = "Detec{\c{c}}{\~a}o de {\'a}reas agr{\'{\i}}colas em tempo
quase real (DATQuaR)",
booktitle = "Anais...",
year = "2015",
editor = "Gherardi, Douglas Francisco Marcolino and Arag{\~a}o, Luiz
Eduardo Oliveira e Cruz de",
pages = "5650--5657",
organization = "Simp{\'o}sio Brasileiro de Sensoriamento Remoto, 17. (SBSR)",
publisher = "Instituto Nacional de Pesquisas Espaciais (INPE)",
address = "S{\~a}o Jos{\'e} dos Campos",
abstract = "Nowadays the challenge in agricultural estimates using remote
sensing is to produce the estimates and data across the crop
season, in near real-time. The aim of this paper is to build an
approach capable to produce the crop maps of soybean+maize in near
real-time, for Rio Grande do Sul state, using MODIS images. To
generate the near real-time crop maps we used the MODIS 16 days
composites vegetation index (VI) images of NDVI and EVI. This new
approach was called Near Real-Time Crop Fields Detection
(DATQuaR). The MODIS VIs images were aggregated in bimonthly
periods using different ways: average, maximum, minimum and median
of registered values. After that, the image of the previous period
was subtracted from the image of the monitored period, generating
the DATQuaR images. These images were classified by slice using as
limit the occupied area estimate with soybean+maize produced by
random sampling over Landsat image and visual interpretation. The
DATQuaR maps were submitted to 3x3 pixel window mode filter. The
results showed that the best approach was to aggregate the maximum
registered MODIS IVs value in the monitored period and the minimum
value registered in the previous period. In this case the EVI
images and the 3x3 pixel window mode filter were used. Using this
approach the DATQuaR method achieved over 81% (in the worst
period, January/February of 2014) of agreement with random
sampling Landsat pixels classified by visual interpretation.",
conference-location = "Jo{\~a}o Pessoa",
conference-year = "25-29 abr. 2015",
isbn = "978-85-17-0076-8",
label = "1146",
language = "pt",
organisation = "Instituto Nacional de Pesquisas Espaciais (INPE)",
ibi = "8JMKD3MGP6W34M/3JM4EF5",
url = "http://urlib.net/ibi/8JMKD3MGP6W34M/3JM4EF5",
targetfile = "p1146.pdf",
type = "Produ{\c{c}}{\~a}o e previs{\~a}o agr{\'{\i}}cola",
urlaccessdate = "28 abr. 2024"
}