For 
with 
r to 
>on- 
(0 
ith 
1, 
Id 
A A SA LO CAREER E DEO. MEE E e me E E E, DNE GEN MER zi ct EE Ep U LE EI A 
ET e iui cen es 
- 6 - 
According to customary usage we define Qu for: 
“= 1 arithmetic mean 
&£ = 2 quadratic-arithmetic mean 
- -1 harmonic mean 
A = -2 quadratic-harmonic mean 
We can now obtain a criterion on the variation of image quality 
over the image field by the ratios: 
-2 
’ go = , 
2 
a; = where qoegject 
ee 
I 
ol 
They lead to a measure of the uniformity of image quality over 
the image field. Again the ratios qj and qo taking into account 
the variation of image quality over the field are unity for 
uniform image quality over the whole image field. qj or qoe<=l 
is therefore not desirable. Usually qj should be adequate because 
q2 is roughly the square of qj and is therefore more sensitive 
to variation of image quality over the field. 
Comparison of image quality of some aerial camera lenses 
Different high quality aerial survey lenses were analysed with the 
proposed method. Averaging of the MTF with respect to spatial 
frequency and field angles led to useful image quality criteria. 
An important aspect of image quality in aerial survey lenses is 
the definition of the appropriate image plane, the choice of which 
depends very much on the averaging method with respect to image 
field. Fig. indicates a typical example of how the choice of the 
best image position depends on the image quality criterion chosen. 
For a typical super-wide angle lens the arithmetic mean Q1 has its 
maximum near the minimum of qj or qo i.e. at positions where the 
MTF over the field varies the most. A compromise seems appropriate 
where the chosen image plane is shifted from the best AWAM (areal 
weighted average modulation) towards a more homogeneous quality 
over the image field (10,11). The chosen image plane is indicated 
by the broken line in Fig.4.