Omitted from all results are: 14 points, eliminated by the president of the study
group; points 16, 23, 71, 185, 247, 260, no sketch available; point 116, outside the
photographic coverage; points 44 and 45, not pricked in the paper prints. This leaves
277 ground control points.
Omitted from the Zeiss Jena results:
points 113, 120, 121, located in a north-south strip only;
point 43, card handling error in the adjustment on corner control;
points 196, 197, 198, 200, card handling error in the transformation of strip 6;
points 21, 22, 46, 87, 91, 227, 252, 267, discrepancies between two sets of
coordinates were considered to be too large;
points 56, 123, 176, 279, residuals larger than 60 m.
Omitted from the Nistri ТАЗ results:
point 216, a large г/-рага11ах;
points 21, 35, 79, 87, 91, 108, 204. 228, 246, 251, 292, discrepancies between
two sets of measurements were considered to be too large;
points 56, 256, 269, residuals around 60 m and larger.
Omitted from the Wild A7 results:
points 113, 120, 121, located in a north-south strip only;
points 54, 75, 87, 114, 193, not measured;
points 163, 200, 245, 276, not punched;
point 3, identification error detected during adjustment;
points 21. 73, 91. 183, 195, 204, 213, 239, discrepancies between two sets of
coordinates were considered to be too large;
point 56, residual larger than 60 m.
d) Analysis of the residuals
The mean square values of the residuals in Table Via have only a limited signi
ficance because of the occurrence of a rather great number of more or less obvious
errors of varying sizes. These errors are mostly identification errors attributable to
ill-defined points or poor sketches and, in the case of the Zeiss Jena and the Nistri
ТАЗ. recording errors.
In order to facilitate the identification of such errors, additional second-degree
conformal transformations have been performed for each strip. For the computation
of transformation formulas, all reliable ground control points have been used.
Points whose transformed coordinates showed an obvious large error have been
discarded, and in all cases where an error was suspected the residuals of the affected
point and of surrounding points have been plotted and compared. Where a residual
shows a discrepancy larger than 10 m with those of surrounding points, the point has
been discarded also. The value of 10 m is rather arbitrary. It has been chosen to
ensure that only points definitely affected by errors were eliminated, even though it is
certain that thus an umber of 0.1 mm errors in the Zeiss Jena measurements, correspond
ing to 8 m, has not been eliminated.