International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004
2.1 Control point network
The earlier control point work for Dione (Davies and
Katayama, 1983) involved image point measurements in 27
Voyager-1 images (1-16 km per pixel) and one Voyager-2
image (5 km per pixel) to determine the latitudes and
longitudes of 126 ground control points. Along with the
2D-coordinates of the control points Davies and Katayama
determined the radius of Dione and defined the position of
the prime meridian.
We re-computed this control point network using a subset of
15 Voyager-1 images with resolutions better than 9
km/pixel and the one Voyager-2 image. Within our
calculations, the position of the prime meridian was fixed at
the value determined by Davies and Katayama (1983).
However, rather than determining the latitudes and
longitudes of the control points only, we solved for the full
3D-coordinates (X, Y, Z) of each control point (Zeitler and
Oberst, 1999: Oberst and Schuster, 2004). This approach
potentially allowed us to determine a higher-order figure of
Dione beyond the sphere.
Over all, there were 135 points (Fig. 1) that were measured in
the 16 images. The total number of image point
measurements was 1482. Image coordinates were converted
to mm on the focal plane using established camera
calibration data (Table 2). Due to the fact that the surface
areas of Dione were mapped at different image resolutions
(see the varying detail across the map) the ability to identify
points was strongly affected. As a result, the distribution of
the points is far from uniform, with a broad point gap to the
west and poor coverage towards the poles.
Using bundle block adjustments, we determined the surface
coordinates of the control points. Image point measurements
and the camera pointing data were introduced as
observations with standard deviations of 9 micro m and |
mrad, respectively. The camera positions were taken fixed at
their nominal values. In addition, Dione's center of figure
was introduced as a ground control point with X=Y=Z=0.
This center of figure was determined from the position of the
limb, which was measured by a limb-fitting program.
5
Liege [7]
Intern
The limb was clearly visible in most Dione images. The X.
Y-. and Z-coordinates of the control points and the three
pointing angles of each image were treated as unknowns and
were solved for.
As a result of the block adjustment we obtained mean point
accuracies of 1.8 km, 2.9 km, and 1.2 km for X, Y, and Z
respectively. The lowest point accuracies are 3.5 km, 6.8 km,
and 2.5 km (X, Y, and Z) and belong to those points located
only within the 5-9 km resolution images (Fig. 2).
K, pix/mm 84.8214
So, pix 500.5
lo, pix 500.5
f, mm VGRI-NAC 1500.19
VGRI-WAC 200.465
VGR2-NAC 1503.49
VGR2-WAC 200.770
Camera parameters are taken from standard navigation
data files (http://naif.jpl.nasa.gov/naif.html) and refer
to the geometrically corrected images
Table 2: Voyager Camera Parameters
2.2 Global Shape Fig. 2
The 3D-coordinates of the control points allowed us to
determine the global shape of Dione. The poor point
coverage of the body (Fig. 1) suggests considering only
very simple body models. Following methods described by
Oberst and Schuster (2004), we performed least-squares fils
to the data using a sphere, a 2-axial ellipsoid, and a 3-axial
ellipsoid, taking into account the radial point errors as
weights in the fitting (Fig. 3). In result we obtained a RM$
value of about 3 km in each case suggesting that a sphere
described the shape of the body sufficient well, and that it
was not meaningful to proceed to higher-order models.
Within this model the radius of Dione was determined to be
R=562.5 +/- 0.2 km consistent with R=560 (RMS=5 km)
determined by Davies and Katayama (1983).
3. DIC
Image |
satelliti
Mimas,
first re,
data de
Howev
taken fi
Fig. 1: Dione base map showing the distribution of the control points.
880
