IEEE Geoscience and Remote Sensing Magazine - September 2014 - 37

TECHNICAL COMMITTEES
GABRIELE MOSER, University of Genoa, DITEN, Italy
DEVIS TUIA, Laboratoire des Systèmes d'Information Géographique, EPFL, Lausanne, Switzerland
MICHAL SHIMONI, Signal and Image Centre, SIC-RMA, Brussels, Belgium

Report on the 2014 IEEE GRSS
Data Fusion Contest: Multiresolution Fusion
of Thermal Hyperspectral and VIS Data

LWIR data were acquired by an 84-channel imager covering the wavelengths
between 7.8 and 11.5 nm, with approximately 1-m spatial resolution. The VIS data
were a mosaic of a series of color images
acquired during separate flight-lines with
approximately 20-cm spatial resolution.
The two data sources cover an urban area
near Thetford Mines in Québec, Canada (see Fig. 1),
and were acquired and provided for the Contest by
Telops Inc. (Canada).
The 2014 Data Fusion Contest was framed as two
parallel competitions: the Classification Contest and
the Paper Contest, as outlined in Fig. 2.
The Classification Contest had the objective of
promoting new methodologies for multimodal and
multiresolution image classification. It ensured a fair
evaluation and comparison of the results obtained
by the different teams. The aim was to provide a land
cover map at the highest resolution into seven land
cover classes (roads, red roofs, grey roofs, concrete
roofs, trees, vegetation, and bare soil). In a first phase,
participants were provided with a subset of the images
and a limited ground truth to train their algorithms.

ach year since 2006, the Image Analysis and Data Fusion Technical Committee (IADF TC, formerly the Data Fusion
Technical Committee) of the Geoscience
and Remote Sensing Society organizes a
data fusion contest aiming at evaluating
existing methodologies and fostering new
ideas at the operational and research level
to solve multisource remote sensing problems [1-7].
I. THE 2014 DATA FUSION CONTEST
The 2014 edition of the Data Fusion Contest provided
a challenging image analysis opportunity, including
multiresolution and multisensor fusion, very high
resolution imagery, and a type of remote sensing data
source previously not seen in past Data Fusion Contests. The dataset involved two image sources acquired
at different spectral ranges and spatial resolutions: a
coarser-resolution long-wave infrared (LWIR, thermal
infrared) hyperspectral data set and fine-resolution
data acquired in the visible (VIS) wavelength range. The
Digital Object Identifier 10.1109/MGRS.2014.2346236
Date of publication: 17 October 2014

Visible

NIR

SWIR

MWIR

Thermal Infrared

FIGURE 1. Details of the multisensor and multiresolution data set for the 2014 Data Fusion Contest: finer-resolution data from a
color camera (left) and coarser-resolution datacube from a thermal hyperspectral camera (right). Data acquired and provided by
Telops Inc.
SEPTEMBER 2014

IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE

Table of Contents for the Digital Edition of IEEE Geoscience and Remote Sensing Magazine - September 2014