CALCULATION OF A BLOCK OF STRIPS, ETC.
9
great computers. We therefore deemed it convenient to elaborate a simplified
procedure, which — whilst keeping practically unchanged the advantages and the
accuracy — would allow the necessary calculations to be performed by means of
small electronic computers (I.B.M. 650 or 1620 type), and in the simpler cases
even by manual calculation with common calculating machines.
This procedure is based on the hypothesis that we have at least two known
points (in the three coordinates) for each stretch, that is six known coordinates
in total. In such a case, six of the seven unknown parameters of absolute orientation
and dimension may separately be deduced for each stretch, and one unknown quan
tity only (that, is, a rotation) remains for each stretch to be calculated in block.
This greatly simplifies the problem, reducing the order of the normal system from
7 n to n.
In practice we can work on the following lines :
a) calculation of planimetry, worked out in first approximation separately
for each stretch, by the rigid model method, using the few A 1 known points included
within the stretch (at least two);
b) calculation of altimetry, worked out in first approximation separately
for each stretch, by the rigid model method and neclecting one y 1 rotation, basing
the M l model on two known points A\ , A\ (axis points), and assuming co 1 = o ;
c) block calculation of the y} rotations — which remain the only unknown
quantities of the problem — performed by imposing in the T 1 *- points, common
to several stretches, the equality of elevations deriving from the i, j, ... origins ;
and eventually the equality to the ground measured elevations in other known
points ;
d) correction of elevations of the T l points for the yj rotations ;
e) calculation of the mean values, from the various origins, of the coordinates
obtained, as stated above, for the T^ - points common to several stretches ;
f) final calculation of planimetry and altimetry for each stretch sepa
rately, by the rigid model method, using now as known points both the A i and the
7'u- points, the latter with their mean plani-altimetric values.
10. — As it can immediately be seen from the above, the simplified procedure
is based on the following hypotheses and considerations :
a) that two known points at least should exist in each stretch. This may
always be obtained by modifying the subdivision into stretches — which is arbi
trary within certain limits — according to the known points available, so that
possibly one of them should be included in the first pair and one in the last pair
of each stretch :
b) that the calculation of planimetry (orientation and dimension), even if
based on two known points only, should give results very close to actual truth.
This will always take place, provided that we respect the conditions usually
required in any aerotriangulation method, i. e. that both ground operations and