IEEE Geoscience and Remote Sensing Magazine - September 2015 - 131

III. THE SUMMER SCHOOL SERIES
While this situation was unfolding, in late 2010 the University of Pavia issued an internal call for International
Summer/Winter School proposals, sized at least one full
week, and including a preliminary selection of the perspective attendees.
It appeared sensible to submit a proposal for a "Data
Fusion in Aerospace" Summer School, linked to the Aerospace Cluster plans cited above. The proposal focussed on
computing performances in the processing of data generated by airborne and spaceborne sensors, and it ranked
first in the assessment at the University level. This result
effectively triggered the School series. The proposal
involved inviting experts from academia, research centres,
industry, and was envisaged as an "education action" of the
Lombardy Aerospace Industry cluster, which contributed
in-kind by entrusting industry experts to teach lectures. In
the following years, favorable combinations of conditions
made it possible to organize yearly editions despite initial
funding was only occasional. In the following subsections,
a brief overview is given of the various editions. Common
features to all the editions are:
◗ The addressed level is PhD, last-year MSc, early PostDoc; professionals in the field were also considered and
should in general also benefit;
◗ An attendance certificate was issued to each student successfully completing the course. No final examination
was planned. In the first edition, the school offered to
issue an additional certificate for completing a choice
of exercises back home. This experiment did not raise
much enthusiasm and was not repeated.
◗ The Schools were advertised through dissemination of
PDF (Portable Document Format) calls on related mailing lists (including at the national level) and publication
of news on suitable web sites (e.g. EOportal). No paper
prints of advertisement material were produced.
A. The 2011 ediTion:
high performAnce compuTing
The opening edition in 2011 was termed "Summer School
on Data fusion and high-performance computing onboard
aircrafts". In 2011 the ESA grid computing system G-POD
[5] had been running for about 6 years; its 2.0 edition was
to be issued at the beginning of the year, offering new features plus access to several Earth Observation datasets.
This prompted reasoning on grid and parallel computing
in remote sensing and aerospace environments. The topic
seemed to be of sufficiently broad interest, and at the same
time to be sufficiently open to applicative use, to hopefully
attract professionals in the field. The main points of the
school were selected to be:
◗ fusion of non-homogenous data;
◗ high-performance computing;
◗ tools available to build efficient solutions to the
above fusion problems in an operational and applicative environment.
september 2015

ieee Geoscience and remote sensing magazine

The above issues are typically faced in modern aircrafts,
whose sensors are capable of providing large amounts of data
of different kinds; these data can be turned into useful realtime information to the pilot or to the ground control stations,
provided that suitable processing is carried out. This is a tricky
task, requiring on the one side to develop effective data fusion
strategies, and on the other to implement them at a sufficient
speed to timely deliver results. Experts of data fusion (see Bordogna [6]) and high performance computing in Remote Sensing (see Plaza [7] and Leporati [8]) were invited to give lectures on the respective fields of expertise. An overview of the
ESA Grid Computing system was also given (see Cossu [9]) to
link with the initial driver of the school. A half-day visit was
organized to Agusta Westland industry premises in Samarate,
Italy. It is to be noted that this first edition of the school was
funded by the University of Pavia itself with a modest contribution (10 k€), assigned after the competitive examination
mentioned at the beginning of this chapter.
Thanks to the University contribution, registration to
the School was free of charge; 26 applications were submitted, all deemed acceptable. 24 individuals attended the
School, of which roughly 62% from Italy, the rest evenly
split between EU and non-EU countries (Israel and China).
Roughly, half of the attendees reported a University affiliation, while the remaining quarters reportedly were from
research centres and companies.
The feedback forms scored on average 4 out of 5 (1 =
worst, 5 = best), and some students congratulated personally to the organizers.
B. The 2012 ediTion: TArgeT deTecTion
The title of the second edition was "Pavia 2012 International School on Data Fusion and Target Detection from
Airborne and Spaceborne Data". While on the previous
year the general problem of efficiently fusing non-homogeneous data was introduced, the step forward of 2012 was
to move towards a specific goal, i.e. detecting "anomalies"
or "targets" in a dataset or data stack. Apparently, this was
a hot topic at that time: according to the scientific indexing
site SCOPUS, there has been a sharp increase in the yearly
number of journal papers published with keywords "data
fusion" and "target detection" between 2009 and 2011.
Some of the main points included:
◗ problems related to target detection in various types of
airborne and spaceborne remotely sensed data, including relevant and efficient data fusion approaches;
◗ available approaches to build efficient solutions for such
problems in an operational and applicative environment;
◗ high-performance computing as a tool to tackle the
-sometimes high- computational burden of target detection algorithms.
Target detection and data fusion issues are also typical
in modern aircrafts, and similar considerations applied as
for the previous edition.
The expenses were partly covered by an ongoing project from a local foundation (Fondazione Cariplo) on
131

Table of Contents for the Digital Edition of IEEE Geoscience and Remote Sensing Magazine - September 2015