IEEE Geoscience and Remote Sensing Magazine - September 2019 - 110

SOFTWARE AND DATA SETS
NATHALIA MENINI, ALEXANDRE E. ALMEIDA, RUBENS LAMPARELLI, GUERRIC LE MAIRE,
RAFAEL S. OLIVEIRA, JAN VERBESSELT, MARINA HIROTA, AND RICARDO DA S. TORRES

Tucumã
A toolbox for spatiotemporal remote sensing image analysis

R

emote sensing images have
been broadly employed
Toolbox
over the past decades to detect and investigate temporal
Purpose
Outcomes
changes on Earth's surface.
Download Worldwide Landsat Pixel Time Series Are
Appropriate tools are needed
Downloaded as Tabular
Time Series Either as Single
Data
to support complex analyses
Data and Images as Raster
Pixels or Entire Images
Acquisition
of temporal data in various
Files.
Manager
domains, ranging from purely scientific to educational.
Purpose
Outcomes
There are many existing tools,
Perform Spatiotemporal
Detect Similar Regions
but it is often necessar y to
Exploratory Analyses and
Based on Their Temporal
Time Series
Generate Similarity and
Profile Similarities
switch among them several
Retriever
Cluster Maps
times. Typical interoperability issues include system mismatch and the variety of user
Purpose
Outcomes
profiles and data types. In this
Syntax Tree of the
Generate GP-Based
Generated GP Individuals
column, we address these isDissimilarity Functions to
GP
and Performance Metrics
Perform Binary Classification
sues by introducing Tucumã,
Classifier
of the Classifier
on Images
an effective open source toolbox for remote sensing image
Purpose
Outcomes
time-series analysis (https://
g it hub.com /n at h me n i n i /
Time-Series Data and
Quickly Explore Worldwide
Landsat Historical Time Series Change-Detection Outputs
tucuma-toolbox). The toolbox
BFAST
May Be Downloaded as
and Apply Change-Detection
is composed of four main
Explorer
Raw Data or Plots
Algorithms
components (see Figure 1):
1) data acquisition manager
(DAM), which supports timeseries retrieval based on se- FIGURE 1. The toolbox overview.
quences of images
2) time-series retriever (TSR), a tool to support timecovery of time-series dissimilarity functions for biseries retrieval based on time-series dissimilarnary classification problems, proposed in [15]
ity functions
4) a recently proposed time-series analysis tool based
on the breaks for additive season and trend (BFAST)
3) genetic programming-based classifier (GPC), a tool
method, called BFAST explorer (BE).
that implements a GP framework to support the disWe present the functional design of the whole toolbox
and common usage scenarios. In particular, we
Digital Object Identifier 10.1109/MGRS.2019.2928260
demonstrate the flexibility of the tool by describing its
Date of publication: 23 September 2019

110

IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE

SEPTEMBER 2019

http://www.github.com/nathmenini/tucuma-toolbox http://www.github.com/nathmenini/tucuma-toolbox http://www.github.com/nathmenini/tucuma-toolbox

IEEE Geoscience and Remote Sensing Magazine - September 2019

Table of Contents for the Digital Edition of IEEE Geoscience and Remote Sensing Magazine - September 2019