considered, to some degree, to be characteristic for the instruments and 
techniques used. 
The information submitted was checked against the stereomodel 
formed on the A7. For this purpose the available information composed of 
numerical profile data (Kelsh K320 and Wild PPO-8), numerical terrain 
information (Gestalt Photo Mapper II), graphical profile information 
(Zeiss GZ-1), and the profile information contained in drop line charts 
(Galileo Orthophoto Simplex, Zeiss Jena Topocart/Orthophot and Ortho SFOM 
9300), was prepared for display at the correct scale on the A7 plotting 
table so that the test measurements could be carried out on the A7 
Autograph. In the case of the Zeiss GZ-1 profiles, recorded on glass, 
this required meticulous measurements on the NRC Monocomparator and sub- 
sequently converting these into the correct profile points at the 
desirable scale. 
As mentioned, excluding major identification errors of the 
profile points, this approach offers an accuracy of the order of +0.2 m. 
The results of this analysis are listed in Table VI. 
It is interesting to compare the standard errors of the recorded 
profile data in Table VI with the corresponding values in Table IV. Table 
IV lists planimetric errors in the orthophotos, determined for points 
located near the scanned profile, which depend largely on the profiling 
accuracy. As expected, instruments with larger elevation (profiling) 
errors also demonstrate larger planimetric errors. 
It is worth noting that a satisfactory height accuracy is 
provided by the automatic image correlation in the Gestalt Photo Mapper 
II. It should be mentioned, however, that terrain points located within 
10 m (1 mm in the original photographs) from larger vertical objects, such 
as groups of trees or buildings, have been excluded in the computation of 
the standard error of the Gestalt height information. In these locations 
larger elevation errors, corresponding to the value AZ in Table VI, 
were determined. max