As for the heights, after the spatial transformation and curvature correction, a
remarkable result was found. It appeared that all strips showed a systematic torsion,
with average values as indicated in diagram No. 3. This might, for instance, be caused
by asymmetrical lens-distortion. (The plate coordinates having only been corrected
for symmetrical lens-distortion, according to the distortion graphs provided by the
I.G.N.).
o
o
o
o
+SÛ /neóers
+//0/n
Diagram 3
Because these discrepancies were too big to directly apply the I.T.C.-Jerie height
adjustment, the heights were first subjected to the Ackermann method of strip-
adjustment [3] (while using all available height control per strip, thus in general the
6 heights as diagrammed in figure 3), executed in the Zebra.
Difficulties encountered: As already reported on July, 26th, 1961, the photographic
quality of the glass-positives was rather mediocre and the identification of many
terrestrial points (the croquis) was poor. During the observations in the Stereo-com
parator, the following code has been used to describe the identification of the ter
restrial points:
1 = good
2 = mediocre
3 = poor
It appears that the observer has coded 93 points as « good », 90 points as « me
diocre » and 109 points as « poor ». During the preparation and marking of the glass-
positives, another photogrammetrist had coded his judgment of the identification
independently: 55 points were coded as «good», 166 points as «mediocre» and 71
points as « poor ».
From the later block-adjustment of the planimetry, it was found that the following
check-points had large coordinate differences between adjacent sections: 2, 12, 28,
39, 46, 48, 49, 76, 81, 104, 114, 119, 156, 158, 167, 168, 182, 204, 207, 211, 215, 221,
236, 239, 246, 248, 254, 262, 273 and 288 (30 points). All, except one, of these 30
points had been coded by the observer with the number 3 (poor).
Obviously, the large discrepancies should be attributed to the bad identification.
In the later block-adjustment of heights, the points 79 and 158 had largely
differing values in the two strips in which they were observed. Point 176 had largely
differing values in the two models of the same strip, in which it appeared.
3. Before the block-adjustment was carried out, an analysis was made of the coor
dinate differences between identical points, situated in the overlap of successive models