- Fiducial centre derived from corner fiducial marks 
- Fiducial centre derived from midside fiductal marks 
- Principal point of autocollimation 
- Principal point of best symmetry 
There are many different coordinate systems in use today by various camera calibration laboratories and 
camera manufacturers (Ziemann 1978 a,b). Because of the different methods, format area on which 
results were based and diversity of reporting calibration results, it is impossible to unambiguously 
interpret or arrive at true comparisons. The reports of the U.S. Geological Survey and of the National 
Research Council may serve here as two examples. At the USGS, the standard practice is the use of the 
principal point of autocollimation as the origin in camera calibration reports and analytical 
photogrammetry programs. For USGS calibration reports, the camera data strip area is used as the 
standard reference on which the fiducial coordinate system is based. At NRC, the principal point of 
autocollimation and the principal point of best symmetry are reported as coordinate differences with 
reference to the fiducial centre (only four fiducial marks are required in Canada), while the location 
of the fiducial marks is given by the six distances between opposite pairs of fiducial marks. A change 
in the reporting of these locations and of the locations of the two principal points and the fiducial 
centre is planned; all points will be reported in a coordinate system with the fiducial centre as 
origin. The x-axis will point to the following photograph (see Figures 2, 4 and 5) disregarding the | 
location of the data strip. Fig. 3 shows the coordinate system for RCl1O cameras as used at the USGS, 
the coordinate system to be used at NRC will be rotated by 180°. Figures 2 through 5 show the 
orientation of the image coordinate system adopted by the U.S. Geological Survey for four different 
types of cameras in general use today. These cameras could be equipped with four or eight fiducial 
marks. The Zeiss-Jena cameras have glass scales with crosses at 10-millimeter intervals located along 
the four sides of the format (external reseau). The end marks of these scales can serve in place of 
the regular corner fiducial marks. 
The determination of the lens distortion for the four semi-diagonals only must be considered as a 
minimum; certain calibration procedures will provide the determination in additional radii, for example 
the USGS procedure, others will result in a distribution of data points throughout the format which are 
not located on a limited number of radii. Since numerical photogrammetric procedures require the 
determination of the overall lens distortion - not only of the radial component - data points for the 
determination of the lens distortion should be distributed such that the lens distortion can be derived 
for the entire format area without extensive interpolation. 
  
  
N A e 25 " | 
— Platen 
identification Y 
(1 & 
Qu) & 
EL X 8-66 
— Lens 
identification 
A FL 
2 oO 
  
  
  
  
n 
' 
eo 
^ 
À 
jM 
  
Fig. 2 Wild Heerbrugg RC8 
The camera is viewed from the back, or a contact positive with the emulsion up. The data strip is 
to the left. The platen identification marker is in the upper left corner. For this orientation the 
film transport is from right to left. There is no asymmetrically located obstruction inside the frame 
area for this camera, however the midside fiducials have independent features that allow for positive 
orientation identification. 
99