EXPERIMENTAL TESTS ON THE BENEFITS OF A MORE RIGOROUS MODEL IN
IMU/GPS SYSTEM CALIBRATION
L. Pinto* , G. Forlani^, D. Passoni
“ DIIAR, Politecnico di Milano, 20133 MILANO, Italy — (livio.pinto, daniele.passoni) (Qpolimi.it
? Dept. of Civil Eng., Parma University, Viale delle Scienze, 43100 PARMA, Italy — gianfranco.forlani@unipr.it
Commission III, WG HI/1
KEY WORDS: Orientation, System, Calibration, GPS/INS, Integration, Adjustment.
ABSTRACT:
System calibration is required in integrated IMU/GPS systems to account for the spatial offset and misalignment between IMU, GPS
and camera frames; synchronization is to be maintained, to predict IMU/GPS position and orientation data at the mid-exposure time
of the images. To this aim, measurement on the ground, complemented by a calibration flight over a test field, are performed.
Depending on the mathematical model, a two steps or a single step procedure may be used to recover the calibration parameters. In
the former, parameters are computed as a weighted average of the discrepancies between the exterior orientation (EO) parameters of
the images derived from block adjustment and those computed from the IMU/GPS data, in the latter parameters are explicitly
inserted in the collinearity equations and the IMU/GPS data are considered as pseudo-observed quantities, replacing EO parameters
as unknowns in the block adjustment. Results of their application to a calibration test field over the town of Pavia are discussed.
1. INTRODUCTION
1.1 A national research program on IMU/GPS systems
Integrated on-board positioning and orientation systems
composed by an Inertial Measurement Unit (IMU) and one or
more GPS receivers allow direct georeferencing of images in
aerial photogrammetry. For sale since a few years, their
acceptance in map production is still hindered in Italy by the
lack of technical prescriptions, required for their use in projects
financed by state or regional map agencies. In spring 2003,
under a national research project aiming at preparing technical
prescriptions and guidelines for use of IMU/GPS systems in
aerial photogrammetry, 4 flights have been executed over a test
field established in Pavia (Italy). Scanned images as well as
GPS and IMU processed data were later distributed to the
research units, all coming from institutes of photogrammetry in
eight Italian universities. Data processing is still underway, to
measure the large amount of images acquired and to perform
the various accuracy checks scheduled within the project goals.
The end of the of data analysis is foreseen for October 2004.
Although every unit may focus also on specific objectives, the
common project goals were set as follows:
=» determination of the accuracy and long term stability of
the GPS/IMU system calibration parameters;
=» development of methods to eliminate residual parallaxes
without resorting to aerial triangulation;
=» development of internal quality control methods in order to
check the reliability of GPS/IMU data, without resorting to
aerial triangulation;
= analysis of the main commercial photogrammetric
software in order to check their functionality related to
direct georeferencing with GPS/IMU data;
= proposal of guidelines for the execution and testing of
airborne photogrammetric flights using on-board GPS/INS
systems.
The specific goal of the research unit the authors are members
of and the subject of this paper is the development of
calibration methods for the integrated IMU/GPS system. More
information on the test characteristics and goals are presented in
(Casella and Galetto, 2003); in the following, a short
description is given anyway to clarify the presentation of the
results of the calibration methods.
1.2 Characteristics of the test blocks
Four different flights have been performed with an Applanix
POS/AV 510 over the test site by Compagnia Generale
Ripreseaeree, based in Parma (Italy). In the first two, the Wild
RC30 camera mounted a focal length of 300 mm, while in the
others a 150 mm lens was used. Each flight resulted one or
more blocks, flown at different heights: image scale are 1:5000,
1:8000 and 1:18000, i.e. the image scales normally used in Italy
to produce maps respectively at the scales 1:1000, 1:2000 and
1:10000.
Flights 1 and 2 have been executed two days apart with a 150
mm lens. They are composed of three blocks each. The 1:5000
block (about 140 images overall) has three ordinary parallel
strips covering a part of the test site, flown in East-West
direction. The first strip, once completed, is immediately flown
in reverse. There are two cross strips, at the head and tail of the
block; each of them is flown in reverse at the end. The forward
overlap is 60%, while the sidelap is 30%.
The 1:8000 block (about 160 images) has seven ordinary
parallel strips covering the whole test site, flown in the East-
West direction. The first one is flown back and forth. There are
two cross strips, at the head and tail of the block; each of them
is flown back (at the end) and forth (at the beginning). The
across-track overlapping is 6096, as well as the across-track.
The 1:18000 block (about 20 images) has a very simple
structure and is constituted by two strips flown in the East- West
direction, with 60/60 overlap.
Internat
Flights
300 mm
but the
practica
Flight 3
structur
1:8000 |
The im:
been sc:
1.3. Te
A test |
Surround
and na
with GP
on verti
eliminat
twice sc
network
The AC
metal p
homoge
chosen
1:5000
paved rc
A small
1.4 Re
The res
measure
1:8000 :
strips, 7
The cro:
one of !
calibrati
manual
intercha
calibrati
300 mm
block fo
points h
(Geosof
154 AG
have be
overall 1
The ref
tangenti
meridiar
calibrati
the prog
Pinto, 1‘
In order
solution.
been use
edges ar
ground
paramet
Fixing tl
compute
164 ava
determin
statistics