nbul 2004
her flight
ssification
sensors is
Megapixel
t area is
sing. And
red to the
ricing the
> classical
h a high
10st in the
ined with
rientation
mponents
using the
POS/AV,
> than the
offers the
eferenced
around of
for digital
he digital
cquisition
sition. the
ed optics
ension of
lual pixel
9 = B/ b o
ng and is
ect space.
luence of
From the
accuracy
parallaxe
;pectively
for high
-to-height
c cameras
) or 0.61
assuming
e, similar
he digital
ie virtual
at matrix
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part BS. Istanbul 2004
arrays, this results in smaller 9 values and less accurate object
point accuracy. Additionally, if digital sensors with non-
quadratic arrays are used for airborne imaging the orientation of
the sensor relative to the aircrafts flight direction influences the
base-to-height ratio. Since the base-to-height ratio is given by
9-1
HRS A008
where s' depicts the sensors extension in flight direction and
p is the forward overlap in percent, the linear dependency of
J and the sensors size is obvious. For example, if the
dimension ratio of the sensor is 4:3 (i.e. 5400 x 4100 pix) the
base-to-height values differ between 2596 dependent on the
sensor orientation. In order to obtain larger swath width quite
often the larger sensor side is oriented perpendicular to flight
direction which again influences the quality of photogrammetric
point reconstruction. This is even valid for the large format
digital sensors (UltracamD, DMC). The DIMAC system
(DIMAC 2004) is differen: Due to the modular design the
orientation of individual cameras heads relative to aircraft flight
direction can be chosen application dependent. If the image data
are mainly used for orthoimage generation based on given DTM
the influence of base-to-height ratio on geometric accuracy is
not as relevant.
3.2 Influence of image motion
Since airborne images are recorded from moving platforms like
aircrafts or helicopters the movement of the sensors during
exposure is of certain influence of the quality/sharpness of the
acquired imagery. This so-called image motion which results in
certain image blur is well known from the classical airborne
cameras. In order to fully exploit the theoretical resolution
power of photogrammetric high performance films forward
motion compensation is realized by physically moving the
photo film during the image exposure. This movement is
synchronized with the mean forward velocity of the aircraft.
Additional rotational movements are compensated from the
stabilized mount which is typically activated during data
recording. Such full motion compensation (translation plus
rotation component) not only allows for a significant Jump in
image quality but also in an extension of maximum exposure
times which results in use of higher resolution films even under
non optimal flight conditions. This increases the number of
potential flight days. Although a full motion compensation is
applied, not all sources of image movements are compensated:
Reasons for remaining image blur are due to deviations from the
assumed mean flight velocity, variation of terrain heights
resulting in a non constant velocity-height ratio and remaining
rotational influences due to the latency of the closed loop
control of the active stabilized mount.
Generally this initial situation is also valid for digital airborne
frame sensors. Two main differences have to be taken into
account: At least for the small to medium format sensors an
active mount is typically not available - this situation might be
different if the cameras are used as sub-system in combination
with laser scanners mounted on a common platform which is
stabilized then. Hence the rotational components of image
motion could not be compensated. The translation effect of
forward motion compensation has to be solved digitally by
moving the charges on the matrix area itself (so-called time
delayed integration TDI), which is typically not available for the
small to medium format digital sensors. In contrary to analogue
film, where the film role is moved in the focal plane of the
camera the digital matrix array is typically fixed in the camera
housing.
For reasons of completeness one specific realization of forward
motion compensation should be mentioned finally: Within the
DIMAC sensor systems, consisting of up to four individual
4080 x 5440 medium format matrix arrays, each array in the
individual camera heads is truly physically shifted during image
recording based on a piezo controlled technique. This approach
is identical to the film based solution. Since there are no
fiducial marks in the digital camera available the movement of
the matrix array has to be known very exactly otherwise the
relation between pixel and image coordinates is not established.
As described above, the medium format digital sensors typically
suffer from influences of non compensated image motion
effects. This gives certain limitations on the realizable image
scales and therefore restrictions on the application fields. The
influence of image motion follows the given equation below:
With the aircraft velocity v, the exposure time / , focal length
c and flying height above ground h, the image motion u is
obtained from well-known formula
u
E Gd vj
2
1
—*y- f . = .
2 h 2 m
where only 50% of the theoretically image motion u, is valid in
UuUx
the images (Kraus, 1990). For analogue imagery the reciprocal
of 1.5 times of the film resolving power is tolerable for image
motion. For digital imagery the influence of motion blur should
be well below one pixel. Since aircraft velocity and image scale
are typically given by default for a certain project, exposure
time is the only variable to minimize effects of image motion.
Exposure time on the other hand is coupled with lens aperture
and film sensitivity given by the ISO value. In digital cameras,
this ISO number is variable over a certain interval which allows
for a larger variation of exposure times, although higher ISO
numbers quite often are associated with higher image noise.
4. MEDIUM FORMAT DIGITAL SENSORS IN
AIRBORNE USE
4.1 Applanix/Emerge DSS
The Applanix/Emerge DSS is one representative of digital
medium format sensor systems. The optical part is based on a
MegaVision 4092 x 4077 pix CCD array digital back mounted
at a Contax 645 medium format film camera housing (Figure 3).
This housing is stabilized using an proprietary exoskeleton in
order to maintain a more or less fixed interior camera geometry.
The camera body itself is rigidly integrated with an Applanix
POS/AV 410 GPS/inertial system providing full high
performance exterior orientation elements for direct
georeferencing. This gives the possibility for fast turn-around
orthoimage generation. In order to obtain regular block
structures (which simplifies the processing significantly) and
active azimuth mount control is realized for the automatic
removal of the aircraft crab angles based on real-time POS/AV
navigation data. The drift correction accuracy is « 0.5deg
(RMS).
