9 
At present this proolem tends to be overcome by introducing standard computer inter 
faces. These may be flexible and thus imply the ability of modifying systems. More 
over, their testing and maintenance is simpler and the cost substantially reduced. 
Another problem concerns the capacity of the computers* storage media and their 
operational speed. At present fast accessible batching storage media tend to be com 
mercially available at acceptable prices. 
Automation in photogrammetry is developing along two basic lines: 
1. For large scale applications phased, sequential operation, with a digital computer 
In a central rôle, appears most adequate. The process can be initiated by procure 
ment of comprehensive digital terrain model (D. T. M. )-data under computer control, 
followed by intermediate editing and conversions of these data, and accomplished by 
automatic contouring, profiling, printing of orthophotographs, etc. 
2. For small scale restitution, systems implementing automatic scanner/sensors, 
correlators, and control elements seem to be appropriate. In such systems digital 
components may perform substantial supporting tasks and/or replace some of the 
automatic analogue components. 
The remainder of this paper will deal with the former line, intended particularly for 
large scale applications. Emphasis will be placed on the acquisition of D. T. M. data 
and the subsequent printing of orthophotographs, as these are two typical photogram- 
metric operations. In addition the concept of digital mono-plotting will be outlined. 
HI. ACQUISITION OF D. T. M. DATA 
The procurement of D. T. M. data is essentially a sampling process which implies 
compression of the original terrain geometry [ 9 ] . Ideally sampling and recon 
struction should be reversal processes. 
The approaches to sampling D. T. M. data may be classified into three categories: 
1. Selective approaches; 
3. Approaches with an inherent data rejection; 
3. Progressive approaches. 
The selective approaches use a pre-determined grid pattern, points selected during 
the sampling process, or a combination of both. Sampling is accomplished in a single 
run and all points sampled are included in the D. T. M. data. These approaches are 
predominantly subjective.