ure 5c,
gistre-
fs sont
mes b
l'obtu-
In voit
re 5d),
ves de
ots de
traité,
re dela
e l'ap-
le à sa
ut dire
e l’ob-
5.
econde
(deux
sifs de
'omme
vitesse
e dans
es plus
le. De
hie en
le pose
econde
t donc
1/1.000
Ilya
e rapi-
r toute
et que
lisible
utilisée
le cette
1X rec-
COMMISSION | Cg)
Appendix to the INT. ARCHIVES OF
PHOTOGRAMMETRY, Vol. XI4, 1956
Differences between Visual and Photographic
Calibrations of Air Survey Cameras™
P. D. CARMAN AND H. BROWN,
National Research Council, Ottawa, Canada
SUMMARY : Significant differences have been found between the calibration
of air survey cameras by visual means and their calibration by a photo-
graphic procedure simulating conditions of use. The differences are
chiefly of chromatic origin.
uE logic of calibrating photogram-
metric cameras by a photographic pro-
cedure was accepted by the Seventh Inter-
national Congress of Photogrammetry in
1952. However, visual methods have con-
tinued in some use for reasons of tradition
and of convenience of equipment. The
Congress agreed that ‘‘a visual method will
be permissible if it... gives the same
values as the . . . photographic method to
within the required accuracy.” Until the
present, both formal and fortuitous com-
parisons of the two methods have shown
no discrepancies which significantly ex-
ceeded the uncertainties of measurement
including those arising from definition
limitations in the cameras tested.
During 1955, tests of Wild RCS5A
cameras on the National Research Council
of Canada’s new photographic calibrator
were found to be showing consistent differ-
ences from the visual calibration data
furnished by the manufacturer. Both the
calibrated focal length and the shape of
the radial distortion curve showed small
but persistent discrepancies. For cali-
brated focal lengths, photographic values
usually exceeded manufacturer's figures
by .01 or .02 mm. Extreme values from the
average radial measured distortion curves
obtained photographically ranged from 10
to 17 microns, averaging 5 or 6 microns
higher than the manufacturer's published
data. The recognition of such small dis-
crepancies was undoubtedly facilitated
by the excellent definition of the Aviogon
lens.
A number of special measurements were
made on two RCSA cameras to explore
the effects of the spectral differences in-
volved in the two calibration methods.
These spectral differences are illustrated
in Figure 1. Curve 4 shows the product
5E., that is the visibility function multi-
plied, wave-length by wave-length, by the
spectral energy curve for CIE illuminant
"C" which is a standardized approxima-
tion to daylight. This is reasonably repre-
sentative of the source-receiver combina-
tions used in visual calibrations. Curve B
shows the product of photographic mean
noon sunlight—that is the standard il-
luminant for photography—the sensitivity
of Aero Super XX emulsion, and the trans-
mission of a Wratten 13 filter. This is a
photographic approximation to the visual
condition of calibration. Curve C shows
the combination used in photographic
calibration. It is the product of mean
noon sunlight, the sensitivity of Aero
Super XX emulsion, and the transmission
of a Wratten minus blue filter.
Results of the special measurements on
the two cameras were essentially identical
hence only one camera will be discussed
here. Chromatic difference of distortion—
also called chromatic difference of magni-
fication or transverse colour—was meas-
ured at a number of field angles. The slit of
a large collimator was illuminated by
various wave-lengths of monochromatic
light from filtered gas discharge lamps and
the corresponding image positions in the
focal plane of the RC5A Aviogon lens
were measured with a microscope. Results
plotted in Figure 2 immediately confirm
that calibration results will be likely to
vary with changes in spectral character-
istics of the calibrating system.
To obtain further information on the
* Contribution from the Division of Applied Physics, National Research Council, Ottawa,
Canada.
623
)F
| X
