Juergen Oberst
Table 2: Angular Coordinates of 5 x 5 grid holes
Id in object space of collimator in degrees
91.63145 89.60811
91.76949 89.60704
91.900728 89.60579
92.04538 89.60463
28 cm 92.18328 89.60350
91.63259 89.74619
91.77063 89.74497
91.90863 89.74376
92.04651 89.74271
92.18443 89.74157
91.63369 89.88405
LY 91.77179 89.88303
"e| 91.90970 89.88176
92.04786 89.88063
92.18571 89.87943
91.63496 90.02232
91.77289 90.02112
91.91097 90.01995
92.04887 90.01881
92.18687 90.01766
91.63613 90.16020
91.77414 90.15910
91.91211 90.15795
92.05009 90.15661 Fig. 2: The calibration grid used for laboratory
92.18815 90.15540 geometric calibration of the MICAS CCD sensor.
{ à hd 2 + +7 eR (C ve] | + i x 1 1
it the holes di i E * t. - 7, f reto d e À i. ar # PU er a
is. Best-fit * T * e. t 8 A + SAM 2 I, 27 e
asurements + ; > Pos s aie I . i Pr s I
shows the at tone dant, a TT et de a LL
model), all 1 e e A, Tu d a ho 1 * «t P T d t *
, with peak a 0 xot T Ï : . * SERES. * oh
nent errors ot BT i Liles AN Taam, Ha, A
the known T D^ TS + 3, c 1 t = ^ ; E rio e
Po v: — 4 S + "e
n (mean of ^ t qu uu s S ra : ; . = À h
wat ^ Moan NAM e Nb m M 4 + u ^o. 1
P Y 4 x + et ut "a Fe) 1 + Ny + + +
, the obj É v 6 mdm” ta
the object LoÓ + 45 1e 4 ë „SO
ved us (not enda «t s id Y i — i wr a "d ; MEME
his type of Ne m E be n ect NE U pedo etu
he nominal pr x T ert ees]: i £T s Bí MN +
o the same s :
"monstrates S
Fig. 3: Vector plots showing offsets between nominal and measured position of grid holes in the 1024 x 1024 image.
Data from 7 grid images (from the total of 21 images used in this analysis) were combined. The plot shows data
residuals before (left) and after (right) the best-fit distortion model (equations 1-3, Table 3) is applied. Note that the
lengths of the arrows are exaggerated by a factor of 50. Hence, the distortions near the image margins before the
correction (left) are as high as 4 pixels.
ear that the
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part Bl. Amsterdam 2000. 223