@InProceedings{FerreiraSantPico:2019:EvDiMe,
author = "Ferreira, Karine Reis and Santos, Lorena Alves dos and Picoli,
Michelle Cristina Ara{\'u}jo",
affiliation = "{Instituto Nacional de Pesquisas Espaciais (INPE)} and {Instituto
Nacional de Pesquisas Espaciais (INPE)} and {Instituto Nacional de
Pesquisas Espaciais (INPE)}",
title = "Evaluating distance measures for image time series clustering in
land use and cover monitoring",
booktitle = "Proceedings...",
year = "2019",
organization = "MaChine Learning for Earth ObservatioN Workshop (MACLEAN)",
abstract = "Time series derived from Earth observation satellite images have
been widely used for land use and cover classification and change
detection. Clustering is a common technique performed to discovery
intrinsic patterns on time series data sets, by grouping similar
time series together based on a certain similarity measure. This
short paper describes an ongoing work on evaluating distance
measures for remote sensing image time series clustering using
Self-Organizing Maps (SOM), specifically to land use and cover
monitoring. We present an experiment to evaluate three similarity
measures, Dynamic Time Warping (DTW), Euclidean (ED) and Manhattan
(MD). In this experiment, we show that ED and ED are more accurate
than DTW for remote sensing image time series clustering in land
use and cover application.",
conference-location = "Wurzburg, Germany",
conference-year = "20 Sept.",
issn = "16130073",
language = "en",
urlaccessdate = "27 abr. 2024"
}