mx = £0.15
the l-order
| times less
1ultaneously
) opposite
still smaller:
s=+1.5 um.
je |-order
tion
ontrol points
ints)
y mp
ais | pixels
2 0.19
)5 0.15
)7 0.19
)8 0.23
)7 0.14
)7 0.13
)7 0.14
)5 0.10
)5 0.09
)9 0.18
)6 0.11
)8 0.17
)7 0.17
)5 0.13
2 0.20
)4 0.12
)6 0.13
)6 0.13
)9 0.23
)5 0.14
)5 0.12
Bt 011
7 0.15
02 | 10.04
Such accuracies are already on the reseau pattern level
of precision. After the correction of residuals dd, the
proportions between errors in X and Y directions
correspond to the optical resolution of scanner.
6. CLOSING REMARKS
1. The UMAX 1200 SE scanner which belongs to the
group of low cost equipment, produces digital images
of rather low geometrical accuracy on the level of
twelve pixels, which can not be used directly (without
calibrations) in digital analytical photogrammetry.
2. That scanner errors are of systematic errors type.
3. After the average systematic errors correction function
(the l-order correction), has been determined there is
possible to reduce the digital image errors to the level
my = £0.15 pixels = +3 um, my= £0.33 pixels = +7 um
4. The l-order image calibration with the use of the
steady surface of mean discrepancies requires only 1
control target on the scanner plate.
5. To introduce to a digital image the ll-order corrections,
which concern residual scanning errors of known
rules of variation, but which must be determined for
each scanner run separately, one must use at least
one row of control targets at the side of the scanner
plate, which can be registered together with the
scanned picture at its side. More reliable corrections
are received with the use of two rows of control
targets embracing the scanned picture on its both
sides.
6. The highest accuracy of correction of geometric
scanning errors was received with the use of the |-
order correction and the ll-order dy, corrections
determined at the presence of two target rows. At that
25
case the following corrected image residual mean
square errors were noticed: my = +0.13 pixels = +2.7
um, my = +0.07pixels = £1.5 um. Such an accuracy is
already on the level of the measurement accuracy of
the scanner reseau control grid used for the accuracy
evaluation.
BIBLIOGRAPHY
Boron A., 1994. Wplyw skanowania na geometrie
obrazów cyfrowych. Archiwum Fotogrametrii, Karto-
grafii i Teledetekcji, Kraków. Vol.1. pp.9.1-9.12.
Boron A., 1995. Homogenity of geometry of images
scanned using UMAX 1200 SE. Polish Academy of
Sciences - The Kraków Section. Proceedings of the
Geodesy and Environmental Engineering Commi-
ssion, Geodesy 38, 1995, pp.133-146.
Jachimski J., 1995. Video Stereo Digitizer - a small
digital stereophotogrammetric working station for the
needs of SIT and other application. Polish Academy
of Sciences - The Kraków Section. Proceedings of the
Geodesy and Environmental Engineering Commi-
ssion, Geodesy 38, 1995, pp.71-93.
Trocha W., 1995. The automated reseau pattern
measurement for the examination of scanned images
geometry. Polish Academy of Sciences - The Kraków
Section. Proceedings of the Geodesy and Environ-
mental Engineering Commission, Geodesy 38, 1995,
pp.147-152.
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B1. Vienna 1996
