New Directions 
It is possible that the problem of automation in photo interpre- 
tation is not fully understood particularly from the viewpoint of the 
human photo interpreters. Until now, most effort has been spent on automatic 
enhancement and pattern änd target recognition by asking the machine to 
correlate the apparent pattern with a stored library. Results have not been 
very promising either from the viewpoint of the researcher or that of the 
user. 
May I suggest, therefore, that more effort be devoted toward pro- 
viding the operator with the means for achieving an enhanced image or record 
and let him decide whether it is more informative than the original document. 
To date, two technologies lend themselves aptly to photographic 
image enhancement. 
1. Electronic - Scanning with video manipulation and computation. 
Optical - Simultaneous operation on large areas of an image. 
Exploring these two technologies, however, in light of recent 
developments we find that significant progress is being made in areas that 
are bound to have implications in the automation of photo interpretation. 
Take the area of electronic scanning for example. Almost all 
previous attempts at photo interpretation automation and image enhancement 
depended upon electronic scanning to convert from the spatial image to a 
temporal signal for processing, manipulation, comparison, and, finally, 
display to the interpreter. The scanning, recording, and display functions 
were generally performed by CRT flying spot scanner techniques. A CRT flying 
Spot scanner was used to scan the input image and convert it into an electronic 
signal which described the point-by-point density values of the photo- 
graph as a function of its light transmission at each point. The signal thus 
15