2. Twelve well distributed control points. Correction of important syste- 
matic errors. 
3. Six control points in the outer borders of the area. Correction of 
important systematic errors. 
In the calculations the tolerance for the exclusion of a control point 
or connection point was set at 0.2 metres. 
Results 
The results can be divided into groups on the basis of the use of the 
calculated coordinates. Errors are defined as the differences between 
photogrammetric and geodetic values. Errors exceeding 0.5 metres radially 
are considered gross errors. These have been excluded from the analysis. 
Of the 75 points measured six had gross errors. 
In the stereo plotting the model seams will give rise to certain discre- 
pancies. These are called connection errors and are calculated from the 
expression: 
i 5 > 
LEN Shen 2 
5, = V X(w + Vy )/n 
in which v. and ^ are coordinate corrections for the po- 
sition of control points and connec- 
tion points 
and n is the number of points. 
The connection error, s, , is a measure of dispersion of the radial stan 
dard error type. It is most closely connected to the term precision. See 
Hallert, 1964. 
Accuracy has been determined with the aid of the errors in the signal- 
led control points which have not been used when calculating transforma- 
tion data. The root mean square of the discrepancies observed is 
/ o 9 
S zr V (c. + B )/n 
m 
in which the discrepancies eo, £, are defined as the difference bet- 
ween photogrammetric and geodetic 
coordinates, x and y respectively, 
and n is the number of points. 
Detailed results are evident in Table 1. 
214