42
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part Bl. Beijing 2008
However, assessment criteria of fusion results obtained from
different sources have so far only been established in a few
cases. Therefore, it is necessary to develop quality standards to
facilitate the comparability of the derived fusion products. Thus,
the aim of this work was to provide an overview of systems,
techniques and applications of sensor and data fusion as well as
to identify algorithms and products for sensor fusion. Moreover,
it had to be investigated whether such approaches are suitable
for standardisation. At the same time, the scientific basis for the
development of a working document for a draft standard for
requirements for geometric fusion methods had to be
established.
This paper results from the standardization of digital
photogrammetric cameras within Germany (German Institute
for Standardization DIN, 2007). However, research into fusion
products needs to be extended with respect to the number of
different sensor types. This is why a comprehensive research
into current photogrammetric and remote sensing sensors was
conducted. The sensors under investigation were systematized
according to their supporting platform (airborne or spacebome)
as well as subdivided in active and passive systems to facilitate
their comparability (see table 1-4). Moreover, the sensor types
were described and their technical specifications were outlined.
Concerning the objective - to extract the information relevant
for standardization - fundamental sensor systems and
approaches were juxtaposed to aid comparison at the end.
The paper is organized as follows. The second chapter contains
an overview of common sensor systems. After a brief
introduction of fusion principles a number of applications for
data fusion obtained from different sensors are considered in
chapter three. Furthermore, some examples of techniques,
methods and procedures for the fusion of geometric data have
been identified, correspondingly arranged and differentiated.
In the fourth chapter some important requirements for fusion
products are derived. Finally, the paper ends with some
conclusions and an outlook.
2. SENSORS
The tables below show an overview of selected sensors (see
table 1-4). The sensors are highly sophisticated and deployed
for production tasks in photogrammetry and remote sensing.
They can be devided into passive and active sensor systems.
Active sensors are able to operate at night. Depending on the
platform (aircraft or satellite) these sensors are classified as
digital cameras, laser scanners, multi- and hyperspectral
scanners and radar systems. This list is not exhaustive. However,
the most relevant sensors from the photogrammetric point of
view are specified. A more comprehensive overview of high-
resolution satellite imaging systems can be found in (Jacobson,
2005).
airborne
spacebome
system
manufacturer
system
manufacturer
ADS40
Leica
Geosystems
IKONOS
GeoEye
DMC
Intergraph
QuickBird-2
Digital Globe
UltraCam
X™
Microsoft
Photo
grammetry
RapidEye
RapidEye AG
Table 1. Passive systems - digital cameras
airborne
spacebome
system
manufacturer
system
manufacturer
AVIRIS
NASA
EnMap
Kayser-
Threde,
GFZ, DLR
HyMap
Integr.
Spectronics
Landsat
NASA, USGS
ARES
DLR
GeoEye1
GeoEye
CASI
ITRES
Table 2. Passive systems - multi- and hyperspectral scanner
airborne
system
manufacturer
ALTM
Optech
ALS50
Leica
LMS
Riegl
LiteMapper
IGI
Table 3. Active systems - laser scanner
airborne
spacebome
system
manufacture
r
system
manufacture
r
E-SAR
DLR
A-SAR
ESA
ARTINO
FGAN
SAR-Lupe
OHB
PAMIR
FHR/FGAN
TerraS AR-X
DLR
AIRSAR
JPL
X-SAR/
SRTM
NGA,
NASA
EMISAR
DCRS
PiSAR
NICT/JAXA
RAMSES
ONERA
Table 4. Active sensor systems - radar
2.1 Digital optical high-resolution sensors
The following approach was taken for quantitative sensor
classification: High-resolution spacebome sensors have a
spatial resolution better than 1 m to 2 m, high-resolution
airborne sensors better than 10 cm to 20 cm. Thus, a resolution
about ten times higher with reference to the Ground Sample
Distance (GSD) is assumed. This is related to typical
applications and sensors deployed in the past. The number of
measured pixels per image exceeds 100 mega pixels. They have
a radiometric dynamic of 12 bit to 14 bit and a signal-to-noise
ratio (SNR) better than 8 bit. Apart from the high-resolution
panchromatic band, optical airborne and spacebome sensors
have another four (mostly lower-resolution) multispectral bands.
2.2 Laserscanner (LIDAR)
The surface is sampled by a scanning laser beam and the period
of time between the emission of the laser impulse and the
receiving of the reflected beams is measured (Time of flight -
TOF). A more accurate approach is based on evaluation of the
phase shift between the emitted and received modulated light
beam. A disadvantage of this approach is the phase unwrapping
procedure (Wehr, 2005).