analyses, they are sometimes regarded with suspicion (Corbett, 1974). It 
would appear, however, that MTF procedures should be judged on the basis 
of results obtained in recent applications (Welch and Halliday, 1973; 
Welch, 1974b; Gerencser, 1976; Gliatti, 1976; Hakkarainen, 1976; Welch, 
1976), of which a few examples based on the author's experience with air- 
craft and satellite photography are summarized in the following paragraphs. 
In an extensive series of tests of photogrammetric camera systems 
good correspondence was obtained between predicted and measured square- 
wave transfer functions (Fig. 6). The standard deviations of the measured 
transfer functions averaged about + 10 percent in response, which is the 
fo Fjord for noticeable differences in image quality (Welch and Halliday, 
1973). 
The steps in the calculation of measured MTF's for the various S-190A 
MPF Skylab camera/film/duplicating film combinations using the EGA techni- 
que are illustrated in Fig. 7. These measured MTF's correspond to within 
6 percent of predicted curves obtained by cascading the MTF of the lens 
with those of the appropriate films and duplicating films; confirming that 
S-190A system performance was about as expected (Welch, 1974b). 
A procedure similar to that employed for the S-190A system evaluations 
also was utilized to assess Skylab S-190B ETC system performance (Fig. 8). 
Component MTF's were cascaded to produce predicted curves, and measured 
MTF's were developed from the edge traces of airfield runway patterns. 
With the exception of the S0-242/2447 film-duplicating film combination 
the edicto and measured values correspond to within 10 percent (Welch, 
1976). 
The agreements between predicted and measured MTF's in these and 
other applications confirm that MTF analysis techniques are appropriate 
for evaluating operational camera system performance (Welch and Halliday, 
1975; Welch, 1975). Resolution estimates developed from these MTF's using 
the procedures indicated in Table II have been equally reliable. 
ELECTRO-OPTICAL SYSTEMS 
The introduction of satellites for remote sensing tasks in the early 
1970's led to the requirement for obtaining imagery of vast areas over 
extended periods of time. A traditional camera system was obviously un- 
suited for this purpose because of limited film supply and the need to 
recover and process the data on a timely basis. As a consequence, electro- 
optical systems employing various mixes of lenses, mirrors, detectors and 
tape recorders were developed. Well-known examples of such systems in- 
clude the return-beam vidicons (RBV's) and multispectral scanners (MSS) 
of LANDSATS-1, -2, and -C (1977) and the conical scanner of Skylab (NASA, 
1976). Other types of electro-optical systems being considered for earth 
survey applications include an improved multispectral scanner (the Thematic 
Mapper) for the LANDSAT Follow-On (formerly EQS/LANDSAT-D), and camera 
systems in which arrays of photodiodes or charge-coupled devices (CCD's) 
replace film in the image plane (Slater, 1974; 1975a; Bisbee, 1975). 
Two parameters which are commonly used to define the performance 
goals of electro-optical systems are the radiometric and spatial resolu- 
tion. Radiometric resolution is expressed as noise equivalent signal (NES)