2 
appropriate coi 
A test target coi 
in Figure 1. T1 
intensity-distribi 
such as sharp-e 
ating image qut 
while the amp] 
and instrumenti 
one-dimensiona 
This paper was first published in The Chartered Surveyor, the journal of the Royal Institution of 
Chartered Surveyors, in January 1964 
Introduction 
URING the past few years improvements in lenses, 
emulsions and cameras have raised the general standard 
of image quality in aerial photography. A questionnaire 
recently circulated by the author revealed a widespread 
opinion that this has increased the accuracy of plotting and 
the ease of interpretation, but that further improvement is 
desirable. This confirms the intuitive feeling that image 
quality is far from the point at which further advances need 
not be sought, but since each step forward now becomes more 
difficult it is very important to have sound evaluation 
techniques. 
Image quality is a very complex matter, but in photo- 
grammetry we are particularly concerned with characteristics 
such as sharpness and graininess which determine micro 
quality and the ability to work at small scales. These are 
influenced primarily by the lens and emulsion but also by 
secondary factors such as image movement, atmospheric 
turbulence, imperfections in filters and windows, etc. If we 
carry right through to the final positive image as seen by the 
plotter or interpreter, we have to take account of numerous 
stages in printing, enlarging or reducing, and the properties 
of the observing instrument, including the human eye. 
Finally, beyond the eye, considered as a physical instrument, 
is the psychological process of recognising and identifying the 
object which corresponds to an imperfect image. The history 
of image-evaluation in aerial photography can be regarded as 
a search for quality criteria which would unify, as far as 
possible, measurements or estimates made at these separate 
and very different stages in a complicated chain. In particular 
there has always been a need for a quality criterion which 
would enable lens and emulsion to be compared in the same 
terms, and for rankings on this scale to be in agreement with 
subjective judgements of picture quality. For some thirty 
years this need has been met by the resolving-power test, 
which in spite of criticism and misunderstanding persists and 
gives the only common quantitative scale for those who 
design and use aerial photographic systems. More recently, 
photographic scientists have taken up the ideas of Fourier 
optics and communication theory, and have applied them 
with some success to the analysis of image structure and 
system performance, as noticed in the author’s paper to the 
1960 Congress. 1 The volume of research has greatly increased 
since that time. Lens transfer functions are now measured 
in many laboratories and are often calculated before the lens 
is made. Emulsion manufacturers supply transfer functions 
for their products. The transfer functions for image move 
ment and other factors are applied in the “ sine-wave 
analysis ” of systems. Surveying the literature, one might 
conclude that great progress has been made and that the 
problems of image-evaluation are nearing solution. How 
ever, it is pertinent to inquire what impact all this activity 
has had on the ordinary practitioner of aerial photogram- 
metry. The answer, confirmed by the questionnaire, is : 
“ Scarcely any.” Sine-wave analysis and the transformation 
of images into spatial frequencies remain almost unknown 
outside the more specialised research institutes and industrial 
groups. Resolving-power remains the standard image- 
evaluation method, widely known and used by practical 
photogrammetrists, although there is a strong feeling among 
some that it is not a reliable guide to the image quality of aerial 
photographs. To the extent that the sine-wave ideas are 
known outside research circles they seem to have led to con 
fusion rather than clarification. Moreover, it cannot be 
maintained that the application of sine-wave techniques is 
essential to the development of better equipment, since some 
of the finest photogrammetric lenses have been designed and 
made by organisations which do not use these methods. 
Current research, in a praiseworthy effort to expose the funda 
mentals of image quality, appears to have got out of touch with 
practical needs. Though this appraisal would be too severe, 
there is enough truth in it to justify a closer examination of 
present trends, pointing out the limitations as well as the 
advantages of the sine-wave approach. Although the 
sine-wave concepts have been very well covered in the 
literature 2 a brief account is given here for orientation and 
easy reference. Experience during the last four years has 
shown that these ideas are not widely known among photo 
grammetrists, and that some repetition is justified. 
Spatial Frequencies and Transfer Functions 
Recent image-evaluation research has recognised that 
directly seen characteristics such as sharpness are the result 
of interaction between more fundamental properties of lenses, 
emulsions, etc. The current “ model ” is developed from the 
conception that the images which we see as individual 
irregularly-shaped tone variations can be mathematically 
represented by summations of sinusoidal wave-trains at 
Figure 1. 
Any object, 
perfect camera, 
determined b> 
simplest specti 
which natural 
actual line oi 
theoretically t( 
falling at a freq 
Since no car 
less changed, 
recorded imag 
of the aerial pi 
tions which Cc 
such as Figur 
as “ modulati< 
normally kept 
may also be e) 
photographic <