International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B6. Istanbul 2004 
Special equipment is used, where all measurements are done in 
very well controlled environmental conditions. The European 
calibrations done for example at the Zeiss (Germany) and Leica 
(Switzerland) calibration facilities based on moving 
collimators, so-called goniometers: The camera axis is fixed, 
pointing horizontal or vertical and the collimator is moving 
around the entrance node of the lenses. The precisely known 
grid crosses from the illuminated master grid mounted in the 
focal plane of the camera are projected through the lens. These 
grid points are coincided with the collimator telescope and the 
corresponding angles in object space are measured. Besides the 
already mentioned calibration facilities other goniometers are 
available for example at DLR Berlin (Germany), Simmons 
Aerofilms in the UK or at FGI in Finland. 
are 
In contrary to the visual goniometer technique, multi- 
collimators are closer to the practical conditions in 
photogrammetry, since the relevant information is presented in 
object space. A fixed array of collimators (typically arranged in 
a fan with well defined angles between the different viewing 
directions) is used, where each collimator projects an image of 
its individual cross hair on a photographic plate fixed in the 
camera focal plane. The coordinates of these crosses (radial 
distances) are measured afterwards and from these observations 
the calibration parameters are obtained. In addition to the 
goniometer method, the multi-collimator is more efficient and 
the calibration includes not only the lens but the photographic 
emulsion on the plate fixed in the camera. Such approach finally 
leads to the more general system driven view — considering not 
only one individual component during calibration (i.e. the lens 
of the tested camera) but including all other important 
components forming the overall system. Although most of 
photogrammetric systems users feel sufficient with the 
traditional system component calibration, the need for overall 
calibration is already obvious since the 1970 as it can be seen 
i.c. from Maier (1978). This system calibration gains in 
importance, especially when including additional sensor like 
GPS/IMU for the data evaluation process. Typically such 
overall system calibrations are only possible with systems in 
situ approaches of calibration. 
3.3 In situ calibration 
In situ calibrations are characteristic for close range 
applications: Camera calibration and object reconstruction is 
done within one process named simultaneous calibration. 
Within this scenario the system and its valid parameters at the 
time of image recording (including all effects from the actual 
environment) are considered in calibration which is different 
from lab calibration described before. Here the camera is 
calibrated in the environmental conditions and at the object to 
be reconstructed. Typically the object reconstruction is the 
primary goal of this measurement campaign, hence the image 
block configuration might be sub-optimal for the calibration 
task. Within other approaches, like test site calibration or self- 
calibration, the calibration is of primary interest. With the use of 
3D terrestrial calibration fields providing a large number of 
signalised points measured automatic or semi-automatic, the 
calibration parameters are estimated. In some cases the 
reference coordinates of the calibration field points are known 
with superior accuracy (test site calibration), although this a 
priori knowledge is not mandatory. Typically, the availability of 
one reference scale factor is sufficient (self-calibration). f 
Since the in situ calibration is a non-aerial approach classically, 
appropriate mathematical calibration models are originally 
developed for terrestrial camera calibration. Substantial 
contributions in this context were given by Brown (1971, 
206 
1966), where physically interpretable and relevant parameters 
like focal length refinement, principal point location, radial and 
de-centring distortion parameters and other image deformations 
are introduced during system calibration. Brown clearly shows 
(from theoretical and practical point of view), that especially 
when using image blocks with strong geometry the method of 
bundle adjustment is a very powerful tool to obtain significant 
self-calibration or additional parameter sets. Such parameter 
sets as proposed by Brown are implemented in commercial 
close-range photogrammetry packages (e.g. Fraser 1997). 
Besides this, calibration in standard aerial triangulation often 
relies on mathematical polynomial approaches as proposed c.g. 
by Ebner (1976) and Grün (1978). In contrary to the parameter 
sets resulting from physical phenomena, such mathematical 
driven polynomials are extending the model of bundle 
adjustment to reduce the residuals in image space. Since high 
correlation between calibration parameters and the estimated 
exterior orientation was already recognized by Brown, the 
Ebner or Grün polynomials are formulated as orthogonal to 
each other and with respect to the exterior orientation elements 
of imagery. Those correlations are especially due to the 
relatively weak geometry of airborne image blocks with their 
almost parallel viewing directions of individual camera stations 
and the normally relatively low percentage of terrain height 
undulations with respect to flying height. In standard airborne 
flight configurations variations in the camera interior 
orientation parameters cannot be estimated as far as no 
additional observations for the camera stations provided by GPS 
or imagery from different flying heights (resulting in different 
image scales) are available. This is of particular interest in case 
of GPS/inertial system calibration due to the strong correlations 
of GPS/inertial position and boresight alignment offsets with 
the exterior orientation of the imaging sensor, which is of 
increasing interest for digital camera systems supplemented 
with GPS/inertial components. Normally, the two modelling 
approaches (physical relevant versus mathematical polynomials) 
are seen in compctition, nonetheless the estimation of physical 
significant parameters and polynomial coefficients is 
supplementary and both models can also be used 
simultaneously, as already pointed out in Brown (1976). 
4. DIGITAL CAMERA CALIBRATION 
Till now only general aspects of camera calibration are recalled 
and very few specifications on the calibration of digital cameras 
were given. Hence, some exemplarily systems already used in 
airborne photogrammetric applications are introduced in the 
following, with special focus on the applied calibration steps. 
Since the individual designs of digital sensor systems are quite 
different, only representatives of the different system classes are 
mentioned in the following, namely the Applanix/Emerge DSS, 
the ZI-Imaging DMC and the Leica ADS40 system. These 
sensors are representatives of the following classes: Sensor 
systems based on (1) 2D matrix arrays within a single camera 
head (typically small to medium sized format) (2) several 2D 
matrix arrays combined within a multi-head solution (utilizing 
medium or larger format matrix arrays for cach individual 
camera head) and finally (3) line scanning systems, where 
several linear CCD lines with different viewing angles and 
different spectral sensitivity are combined in one focal plane. 
The DSS is representing the systems of the first class. This 
group is a very vital one, since many of the already relatively 
low-cost semi-professional or professional digital consumer 
market cameras can be modified for airborne use. Petrie (2003) 
presents a very good overview on the 2D digital sensors market