379 
bandwidth but poor modula- 
1 purposes this test will rank 
l for that purpose the author 
surpassed by any other test, 
isolated detail can be seen, 
: and proven test which will 
the past few years have not 
valuation of a system or the 
ave greatly improved our 
f image quality, so that we 
valuation technique can or 
>GMENT 
f Itek Corporation’s per- 
The opinions expressed are 
:hor, but their development 
evitably been influenced by 
gues at Itek and (formerly) 
iment, and with Professor 
1 programmes sponsored by 
¡ted States Air Force under 
L To all of these grateful 
CES 
i. Special Congress Number B. 
is given by Perrin. Perrin, F. H. 
neers, 69, 151, 239 (1960). 
'ournal of Photographic Science, 10, 
'ourth Engineering Report. Itek 
A.ir Force. 
IMAGE LOCATION 
By R. W. Fish 
Ministry of Aviation 
(Invited Paper, Commission I, Xth International Congress of Photogrammetry, 
Lisbon, 1964) 
Abstract 
An account is given of the source of position errors in air photography, 
with particular emphasis on results reported since the I960 I.S.P. Congress 
in London. Some proposals are made for work to be done. 
1. Introduction 
In the last fifteen years or so there has been a revolution in the optical performance 
of air cameras, largely due to the introduction of the Aviogon lens and its later 
competitors. With an improved resolution right to the edges of a 90° field, it has 
become possible to locate images with greater precision, and gradually the photo- 
grammetrist has come to expect greater accuracy from his negatives. At the 
Symposium of Numerical Photogrammetry in Milan 113 there was considerable 
discussion on the need for greater accuracies in the photography. Residual errors of 
about 20 ¡x were mentioned as being found in “good” photography, while the makers 
of plotting machines have provided a measuring accuracy of perhaps 2 p,. The 
situation is, however, now rather different from any we have met in the past, in that 
there are a large number of possible sources of these residual errors, each of which 
can contribute a few microns. The camera maker must therefore make significant 
improvements in many directions if the total error is to be reduced to 5 /x or less. 
The present paper reviews recent work on these sources of error, concentrating 
on developments reported since the 1960 London Congress. 
2. Basic Geometry of Image Formation 
As in many branches of science, the basic concepts of image formation present 
difficulties in definition when applied to real systems which are far from perfect. 
These difficulties continue to provide opportunities for discussion, but for our 
present purpose it is only necessary to present a simplified account of the subject. 
Image formation is a process which transforms an object point (R, 6) in 
polar coordinates on a plane earth into an image point (r+Ar, 6+ Ad) on a plane 
negative. By definition there is a constant relationship between R and r over the 
entire negative given by 
R = kr = HrIF (1) 
where k is a constant 
H is the aircraft height 
F is known as the principal distance 
Ar is known as the radial distortion 
r/\9 is known as the tangential distortion.