ul 2004
555
ont
+
d. +9
à
Pat
b En
Lu
edge
image
lenging
emplate
tures to
avoids
method
>nted in
tive for
ate and
orm to
11-122.
iniques.
del for
versity,
studyon
und in
^s. Vol.2
ing the
Vision,
n-made
fusion
'ensing,
nection
AN ITALIAN RESEARCH PROJECT ON DIRECT PHOTOGRAMMETRY
R. Galetto, V. Casella, M. Franzini, A. Spalla
DIET — University of Pavia, via Ferrata, 1- 27100 Pavia, Italy
(riccardo.galetto, vittorio.casella, anna.spalla, marica.franzini) @unipv.it
KEY WORDS: Photogrammetry, test, GPS/INS, calibration, triangulation, adjustment, accuracy. standard.
ABSTRACT:
The paper deals with an Italian research Project on direct photogrammetry, active in the period 2003-2004. Among its most interest-
ing features are: usage of a rich test site, having 160 artificial GCPs and 60 natural ones; performing of four dedicated flights, each
containing different blocks at different scales. The goals and the phases of the Project are described, together with the test site and the
several, already performed, ad-hoc flights.
The aim is to favour knowledge exchange and collaboration between Italian researchers and those from other countries.
1. INTRODUCTION
An Italian national research project on inertial positioning in
photogrammetry is in progress. It is chaired by one of the au-
thors, prof. Riccardo Galetto of the University of Pavia and it
was joined by nine units belonging to seven Italian universities.
They are listed together with the units' leaders: Pavia - Profes-
sor Riccardo Galetto; Milan - Professor Carlo Monti; Como -
Professor Fernando Sansó; Torino - Professor Sergio Dequal;
Vercelli - Professor Ambrogio Manzino; Parma - Professor
Gianfranco Forlani; Sassari - Professor Maurizio Minchilli; Pa-
lermo - Professor Benedetto Villa; Perugia - Professor Fabio
Radicioni.
The aim of the project is to investigate almost all the features of
inertial positioning: Kalman filtering of the measurements ac-
quired by the sensors; sensor calibration; precision and accu-
racy; direct sensor integration versus integrated sensor orienta-
tion; stability of calibration; influence of bad GPS constellation
and/or cycle slips; local reduction of parallaxes. In addition,
intrinsic methodologies will be developed and tested for quality
assessment without ground truth. All these investigations will
lead to the definition of guidelines for performing GPS/INS
aided photogrammetric flights.
A qualifying feature of the Project is Pavia’s test site. Among
the infrastructures, we have a levelling network, a high preci-
sion GPS network and a GPS permanent station. There are
more than 100 natural GCPs, which have been measured with
GPS. There are five photogrammetric coverages acquired in the
last three years by planes equipped with inertial sensors. To
carry out the project, 180 artificial GCPs have already been
drawn and measured. Four new dedicated flights were per-
formed: they are characterized by three different image scales,
1:5000, 1:8000 and 1:18000; two flights were performed with
the 150 mm camera and the remaining two were performed
with the 300 mm camera.
2. GOALS OF THE PROJECT
The structure of the Project has been defined after careful ex-
amination of the well known OEEPE project Integrated sensor
orientation.
Between the two Projects there is affinity, concerning the fun-
damental methodology which has been followed, i.c. the use of
891
test sites, each of which has many ground checks measured via
GPS. There is continuity too, because the results of the OEEPE
test have been chosen as a starting point for our Project and its
further aims are to widen the sampling and to solve problems
which have not been tackled in the European one, such as the
possibility to reduce the residual parallaxes locally, the devel-
opment of intrinsic methodologies in order to control the quality
of the results and the drawing up of a series of rules for the
planning and testing of the GPS/IMU assisted photogrammetric
flights.
On the other hand there are also differences between the two
Projects. First of all the data examined in our test will consider
at least two test sites and they will be marked by a broader
range of heterogeneity: for instance cameras with 150 — 300 mm
focal length will be used. Besides that, several photogrammetric
scales will be taken into consideration, chosen from those more
commonly used in Italy for large and medium scale surveys,
1:5000, 1:8000, 1: 18000.
Finally, both artificial and natural ground points will be used as
photogrammetric control points. Artificial ones obviously ob-
tain a better collimation and an enhanced accuracy, which how-
ever is not realistic, that is to say that it is not comparable to that
which can be obtained by common cartographic production.
The utilisation of natural points allows for more realistic estima-
tions of the accuracy level of the direct photogrammetric proc-
ess in its industrial application.
[n order to reach its objectives, the project has been divided into
two phases, each one lasting one year and based on a different
data set.
In the following an itemized synthesis of the Project will be pre-
sented. At the end of each item, bibliographical links are listed
to papers which have been produced within the frame of the
Project and which cover the related topic, at least partly.
2.1 Objectives of the first year of research (2003)
a. Study of the quality of the external orientation of photo-
graphs obtained in an industrial environment (Casella et al.,
2004).
