is evident stems from the thought mentioned earlier that photo inter- 
pretation involves different techniques with different people depending 
upon the required output. This same philosophy now surrounds efforts 
toward automation in photo interpretation. The result in some instances 
is promising and in due time will undoubtedly offer useful and economically 
feasible automation of some previously tiresome, repetitious, time consuming 
aspects of photo interpretation. Quite often, however, concepts are being 
explored to automate interpretive functions that either cannot be successfully 
automated, or, if at all feasible, represent such a cost that they may never 
be warranted from an economic viewpoint. 
Symposiums and Proceedings 
  
There are those that believe that a technology or some branch 
of a science does not warrant a classification of its own until it reaches 
the Symposium status. If this is the case, it can be stated that automation 
in photo interpretation, at least in the United States, does warrant such 
classification since several symposiums and specialized meetings concerning 
photo interpretation automation have been held in the past four years. 
Probably the largest of these was the Symposium on Automatic 
Photo Interpretation held in Washington, D.C., May 31 to June 2, 1967, 
co-sponsored by the Information Systems Branch of the Office of Naval Research, 
the Computer Science Center of the University of Maryland, and the Pattern 
Recognition Society. During the three day session papers were presented 
under 6 categories ranging from Target Recognition to Learning Machines. 
Based upon the papers presented, it appears that the automation 
of conventional aerial photo interpretation as currently practiced is a