of adjacent light and dark squares each having an image dimension of 
approximately 200 x 200 um is the minimum acceptable target size. Ob- 
viously, larger targets will reduce errors in slit alignment and permit 
multiple scans from different sections of the edge to be taken and ave- 
raged. At relatively large scales (e.g. >1:50,000) typical edge targets 
are found among the geometric constructions of man, whereas at very small 
scales (e.g. =1:3,000,000 of LANDSAT and Skylab S-190A Multispectral 
Photographic Facility (MPF) images) reduced ground resolutions of 60- 
250 m permit the use of field boundaries and shorelines. For the analysis 
of satellite images with ground resolutions of 15-30 m for low-contrast 
targets (as produced by the Skylab S-190B Earth Terrain Camera (ETC)) 
the edges formed by large airfield runway patterns provide suitable targets. 
The fidelity of a target recorded by a wide-angle photogrammetric e 
camera system is also influenced by target orientation and format position. 
For example, edges oriented perpendicular to the flight direction will be 
degraded by image motion and those in the corners modified by lens aberrations. 
These problems may be taken into account by noting the orientation of the 
edge with respect to the flight line and selecting edges within approxi- 
mately 10 degrees of the optical axis. Unless an obvious flaw in the 
imagery is apparent upon visual examination, analyses of targets near the 
center of the format will indicate whether or not satisfactory system per- 
formance has been obtained. 
Although techniques relying on edge measurements are excellent for 
analyzing operational imagery, the search for suitable targets can be an 
exceedingly frustrating task. Consequently, a record should be kept of 
the geographic location, size and contrast of edge targets selected for 
evaluation purposes. A filing system can simplify the task of 
analyzing repeat or cyclic missions. 
Degradations Introduced by the Microdensitometer and Film Granularity 
  
Slit size is perhaps the most important parameter governing micro- 
densitometer performance. For photogrammetric photography with maximum 
spatial frequencies of 50 to 60 cy/mm, slits of 2 x 200 um or 3 x 200 um 
can be used without introducing noticeable degradations into the system. 
These apertures minimize the troublesome problems of false or erratic 
response due to coherence and film granularity without resorting to 
digital smoothing functions, and are significantly larger than the ap- 
proximate 1 x 80 um slit frequently used with fine-grained, high-resolu- 
tion images (Gliatti, 1976; Gerencser, 1976). The relationship between 
the microdensitometer MTF for various slit widths (based on (sin x)/ 
x functions) and the MTF of a typical photogrammetric system is shown 
in Fig. 5. Even for a 3 um slit width, microdensitometer response is 
better than 90 percent at the spatial frequency limit of the camera system. 
Photographic System Performance 
  
It is apparent from the preceding discussion that the evaluation of 
system performance based on MTF's requires a skilled analyst with consid- 
erable knowledge of the photographic process and of the factors governing & 
image quality. In fact, a high degree of subjectivity is involved in 
selecting edges and conducting measurements from which the MTF's are 
determined objectively. Because of the complexities involved in MTF