32 
SYMPOSIUM PHOTO INTERPRETATION, DELFT 1962 
calculation the slit image is projected on a test pattern (fig. 3) in two different 
manners. In the first case the test pattern may be located so that as much light 
of the slit image as possible falls on the transparent interspace between the bars 
of the pattern. In the second case the pattern is located in such a manner that 
the slit image mainly falls on the opaque bars, with a minimum of light trans 
mitted through the pattern. Consequently, one maximum and one minimum 
intensity of transmitted light are obtained. Then the contrast is determined as 
the ratio (I max — Imin) : (Imax + Imin). In fact, this is the analogue to the 
modulation in the field of electronics. The result of the calculations that the 
broader the slit image, the smaller is the percentage of contrast transferred. 
The method is based on a sinusoidal test pattern (fig. 4). However, a square 
wave pattern may also be used, if a mathematical transformation is performed. 
The entire procedure implies that the micro-photometer trace of the slit image 
is Fourier-analysed. The explanations of these facts may be beyond the scope 
of this paper. 
The calculation of the transfer curve may be done by hand, but this is fairly 
time consuming - about one hour per frequency. Five or six frequencies are, 
however, sufficient for aerial photographs. The fastest method is to use an 
electronic computer where the transformation from square-wave to sinusoidal 
test pattern is also worked into the programme. 
T-functions for factors affecting the aerial photograph 
In judging an aerial photograph, the interpreter wants to know all the 
factors which affect the quality. The most important ones are the lens, the him, 
the shutter efficiency, the different kinds of vibrations, the image motion, and 
the aerial haze. Here the T-functions offer essential advantages over the older 
methods. In fact each of these factors may be expressed as a T-function, and 
all the functions may be multiplied, frequency by frequency, giving the result 
ant function of the aerial photograph, as pointed out by, among others, 
Scott [9]. 
As shown earlier by the author [5] there are also possibilities of obtaining 
the entire T-function of the aerial photograph directly. This work was carried 
out with a row of fluorescent lights on the ground, photographed at night. 
Since that occasion a simpler manner has been used with satisfying results. 
In addition, this method is more in agreement with the conditions of aerial 
photography in practice. A comparison between this „practical" T-function 
obtained from the air and the function determined by multiplying the T- 
functions for the factors mentioned above will be shown. However, before this 
is done the six factors ought to be discussed, taking as a starting point actual 
photography recently carried out in Sweden. 
The lens-function was measured at the Institute of Optical Research with 
an instrument described by Ingelstam [7]. At first only the image centre 
function was measured. This summer the measurement of “contrast mountains”