S
737.
n con-
r
1S under
) mm
4,000m
nded in
normal
ometers.
s of the
gree for
nslation
entation
control
The
points
of unit
re given
uss the
oach of
| Case is
] angle,
height
luences
gh the
| 2,000
cal one,
maximum height difference : 100m
iteration step I I
standard error of 2.1 2.1
unit weight( um) ^ ,
average internal 8.2 6.2
error(m)
average external 4.6 4.6
error(m)
maximum height difference : 2000m
jteration step I II
standard error of
unit weight( um) 2.1 2.0
average internal 9.0 8.5
error(m)
average external 8.6 8.1
error(m)
maximum height difference : 4000m
iteration step I II
standard error of 2.3 2.0
unit weight(, m) : :
average internal 10.7 8.7
error(m)
ayerage external 15.4 8.5
error(m)
Table-1 : obtained results in simulated
space triangulation using affine
transformation
if the terrain is extremely flat. However, the
external error increases with the height differences in
the terrain.
3) The iterative orientation calculation becomes
effective when the maximum height difference in
the terrain exceeds 4,000 meters. This means that
the proposed correction technique of the image
transformation errors is mathematically sound.
CONCLUDING DISCUSSIONS
This paper has presented an iterative orientation
method of satellite CCD camera imagery based on
affine transformation and developed an effective
correction approach of the image errors which arise in
the transformation of the central-perspective CCD
camera images into affine ones. The proposed
orientation technique has been tested with a simulated
example of 22 satellite CCD camera images taken
consecutively and proved to have a high accuracy.
REFFRENCES
/1/ Okamoto. A. : ULTRA-PRECISE MEA-
SUREMENT USING AFFINE TRANS-
FORMATION. International Archives of
Photogrammetry and Remote Sensing, Vol.
615
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B3. Vienna 1996
29, Commission V, (1992), pp.318-322.
