14 
20 microns, it seems that the standard deviations must lie in the range 
smaller than 0.14 per mil of the flight height. And if there are errors 
larger than this value, it must be resulted by bad identification. Even 
it may be said that sometimes our photogrammetric survey exceeds the ground 
survey in average accuracy, for it appears that in the case of signalized 
identification, the standard deviation of residual errors at controls 
grows better for higher flight height. This means that in the case of 
lower flight height, the average accuracy of photogrammetry might exceed 
that of ground survey. Consequently, the relative errors which are the 
aggregation of photogrammetric errors, errors of ground survey and errors 
of identification have necessarily some limit in the case of lower flight 
height. 
The operation of the electronic computer for absolute orientation is 
the most difficult procedure in our analytical system. Rejection of a point 
not to be used for the computation or reinsertion of the same point, 
selection of equation to be used, all the operations have to be done very 
quickly, because the time lapses before the computer very quickly during 
the operator’s contemplation stopping the computer, and the down time of 
the computer will much reduce the efficiency of the computer. Automatic 
discrimination by computer is very much difficult in this phase, the causes 
of which are mainly the irregular occurrences of mistakes and errors of 
various kinds at control points. It needs ordinarily 10 or 15 minutes 
for the detection of wrong points by absolute orientation of one strip, and 
there is very rarely a case where we find no wrong point. Two or three 
times of recomputations often occur before we attain the final result. 
By simple instruction of the operator which is the signal of finish 
of the computation, computer punches out through the high speed punch 
out device all the coefficients of transformation equation finally used 
for absolute orientation. This out put tape becomes the input tape of 
next program, as already mentioned, with the other tapes which are the 
output of the programs 1,3 and 4. The next program is the final program 
in which all the results necessary for plotting are given completely.