By substitution the equations (1) and (2) may be expressed as follows:
1
ce = 1 (+ 3 Az, —1 App) 6
1
c= 7 (— 2 Azn + l'An) €)
where Az, and Ag; are the errors of the absolute orientation of the last stereoscopic model, i.e. for
x — |. The curve of the 5rd degree goes over into one of the 2nd degree, if the condition
ea 20. (0)
is satisfied, i. e.
2 Az,
Aqy = — 3 % : (8)
In this case
Az, Ag
Gscpgee c. ©)
Similar formulae are valid for the adjustment of Ax and Am, as well as for Ay and Aa. Only o
cannot be adjusted in this way.
Assuming that parallel with this first strip a second strip is flown with a lateral overlap of about
20 per cent, we may adapt the second strip to the first one in a similar way as the second stereo-
scopic model is adapted to the first stereopair within one flight strip. With two identical points at the
beginning and at the end of both these strips, the values dx, dy, dm, da, dz and dg of the second
strip are determined. Only the lateral tilt do remains unknown. On account of the insufficient overlap
of strips this tilt cannot be determined in the same way as the difference do in the longitudinal tilt
with the usual bridging of successive models. For this purpose different possibilities are available. By
repeating the procedure we assemble a series of flight strips into one block. If control points are also
available for the last flight strip an absolute orientation can be performed for that strip as with the
first strip. From the different coordinates of that strip we compute the errors of closure in our block,
and these errors can subsequently be adjusted in the same way as the errors of closure in the last stereo-
pair of a strip.
In this or a similar way two or more strips have already been connected so far. But in this case the border
strips only are correct. Perhaps two transversal strips at the beginning and at the end of the block are
sufficient to furnish additional determinations and consequently better results. Experience has proved
that larger discrepancies occur in all three directions of coordinates at the border lines of the strips.
Especially the elevations nearly always show considerable discrepancies.
As an example the adjustment of y-deviations may be chosen. The differences in bridging two adjacent
strips be
di-1,i = Yi— Yi-1. (10)
The values y; and yi.; are corrected in such a way that
Di-1— Ui — di-1,i. (11)
Thus we only have n — 1 equations for the n unknown corrections v. We now determine the values v
in such a way that
[oo] = min. (12)
Then
Di — 0i-1— di-1,i — 01— d1,2— de,g —...— di-1,i. (15)
With
si= di 2+des +... di-11. (14)