IEEE Geoscience and Remote Sensing Magazine - September 2013 - 37

FIgUrE 1. Composition of the hyperspectral and LiDAR data
sets over the University of Houston campus.

method used, and the final result generated. The papers
submitted were automatically formatted to hide names and
affiliations of the authors to ensure neutrality and impartiality of the reviewing process.
The 2012 Contest was designed to investigate the potential of multi-modal/multi-temporal fusion of very high spatial
resolution imagery in various remote sensing applications [6].
Three different types of data sets (optical, SAR, and LiDAR) over
downtown San Francisco were made available by DigitalGlobe,
Astrium Services, and the United States Geological Survey
(USGS), including QuickBird, WorldView-2, TerraSAR-X, and
LiDAR imagery. The image scenes covered a number of large
buildings, skyscrapers, commercial and industrial structures,
a mixture of community parks and private housing, and highways and bridges. Following the success of the multi-angular
Data Fusion Contest in 2011, each participant was again
required to submit a paper describing in detail the problem
addressed, method used, and final results generated for review.

buildings, highways, railway, and vehicles. The validation
samples that the Contest organizers used to evaluate the submissions were not disclosed.
Best Paper Challenge, with the objective of promoting novel use of hyperspectral and LiDAR data. The deliverable was a 4-page manuscript describing the problem,
methodology, results and discussion. The goal of this challenge was to encourage the participants to consider hyperspectral and LiDAR data fusion problems and to demonstrate novel and effective approaches to address them.
The Data Fusion Award Committee consisted of seven
independent judges from universities and industries:
◗ Jocelyn Chanussot, Grenoble Institute of Technology,
France
◗ Melba Crawford, Purdue University, USA
◗ Jenny (Qian) Du, Mississippi State University, USA
◗ Paolo Gamba, University of Pavia, Italy
◗ Fabio Pacifici, DigitalGlobe, Inc., USA
◗ Antonio Plaza, University of Extremadura, Spain
◗ Saurabh Prasad, University of Houston, USA
Papers were judged in terms of sound scientific reasoning, problem definition, methodology, validation, and
presentation.
3. OUTCOME OF THE CONTEST
More than 900 researchers from universities, national
labs, space agencies, and corporations across the globe registered to the Contest, demonstrating the great interest of
the community in the DFTC activities in promoting cutting-edge research of remote sensing image processing and
analysis. The data sets were downloaded from a total of
69 different countries, with a large number of registrations
from less developed areas. Fig. 2 shows the geographical
distribution of the participants, where other indicates the
sum of all countries with less than 10 participants.

2. 2013 DATA FUSION CONTEST
The 2013 Contest was aimed at exploring the synergetic use
of hyperspectral and LiDAR data. The hyperspectral imagery was composed of 144 spectral bands from 380 to 1050
nm. A co-registered LiDAR derived Digital Surface Model
(DSM) was also made available to all participants. Both data
sets had the same spatial resolution (2.5 m). As shown in
Fig. 1, the data was acquired by the National Science Foundation (NSF)-funded Center for Airborne
Laser Mapping (NCALM) in the summer
of 2012 over the University of Houston
USA
17%
and the neighboring urban area. The data
17%
China
pre-processing was conducted by student
India
volunteers at UH's Hyperspectral Image
Iran
1%
Canada
Analysis group, and NCALM staff. A
1%
1%
France
ground truth corresponding to this data1%
Germany
2%
set was created by the contest organizing
Italy
17%
2%
committee via photo-interpretation.
Pakistan
This year, the Contest consisted of
2%
Spain
two parallel competitions:
Turkey
2%
Best Classification Challenge, with the
Japan
Belgium
objective of promoting innovation in clas2%
Malaysia
sification algorithms, and to provide fair
3%
Egypt
performance comparisons among state3%
Greece
14%
of-the-art algorithms. For this task, users
Other
4%
were provided with training samples from
8%
14 classes of interest, including various
types of vegetation, soil, water, but also FIgUrE 2. Geographical distribution of the participants for countries with more than
less common targets, such as commercial 10 participants.
September 2013

ieee Geoscience and remote sensing magazine

Table of Contents for the Digital Edition of IEEE Geoscience and Remote Sensing Magazine - September 2013