controlling the systematic errors of the photogrammetric process. As 
mentioned above the combination of photogrammetry and dataprocessing 
affords a good posibility of increasing the geometrical quality by adding 
corrections for systematic errors, provided that their mathemathical form 
is known. A great deal of the errors have been investigated, better film- 
bases etc. has also reduced the systematic errors and the photogram- 
metric equipments are more stabile than before. The knowledge of the 
systematic errors is, however, still far from acceptable. The influence of 
the systematic errors can quite obvious be compensated by using a lower 
photographing altitude, but then the economy and the efficiency of 
photogrammetry will be decreased. 
Another problem when using photogrammetry in the final design 
arises from the vegetation, bushes, trees etc. The photographs should be 
taken without leaves on the trees but in some cases problems still arise. 
Therefor it might be an idea to uncover the vegetation before photog- 
raphy. That can however become somewhat complicated as the right of 
way-area is depending as well on the horizontal as on the vertical align- 
ment of the road and some changes may be made due to the detailed data 
of the terrain in the final design. Therefor the computing systems must be 
so designed that as well photogrammetric as terrestrial terrain-data as a 
combination of them will be accepted as input. There must in other 
words be included possibilities of completing the photogrammetric datas 
with terrestrial ones. This requirement has also been met in most of the 
existing system. 
Test works have been carried out in several countries and they have 
clearly shown that the quality of the photogrammetric data is fully 
comparable to that of terrestrial data, when the massquantities are 
discussed. Within the different cross-sections the deviations between 
photogrammetrically measured datas and terrestrial ones may become 
larger than the tolerances of the coordinates. Experiences have 
shown that this depends partly on the difficulties of referring 
photogrammetric cross-sections to the terrestrial ones and partly to the 
fact that the choice of points within a cross-section is essential for the 
shape of the whole section. The system of cross-section is unfortunately 
not a good method for describing a terrain surface, but it is so far the 
principle mostly in use. 
225