'1' i >v‘ 
(172) 
for the lens-film combination. These are shown for the two symbol types: for 
periodic cymbels (the resolution target) and isolated symbols (isolated white 
lines on a black background) (Figure 3). The difference in the response of the 
systems to the two types of detail may well throw doubt in our minds as to 
the validity of the typical resolution test object. 
Let us now assume that in the object space we have an array of symbols 
Fig. 3. Object Contrast Thresholds. 
distributed at random in both dimension and contrast. Thus we assume that 
in this object space an equal number of symbols exist in any given unit area 
(Ac) (Ax). We may select the particular range of sizes (x ma x—x m in) to be 
recorded, for this is a function of the choice of scale. Based on this first as 
sumption we then see that the number of symbols in the object space which 
will be recorded by our system depends upon the area of the object space that 
existe above the threshold of the system. 
But the first assumption is not correct. The fact that aerial photography