Fig. 2: Perspective view from the South of the lunar Humboldtianum basin showing a section of the terrain model from Fig.1. 
The inner and outer basin rings have diameters of approximately 270 and 500 km, respectively. True heights vary by as much 
as 6 km and are exaggerated in this view. Axes have arbitrary units. 
Heights vary from -4500 m to +2000 m. The Humboldtianum 
basin displays a clear double-ring structure difficult to see in 
image data alone (Fig. 2). The basin has a depth of 6000 m 
below the lunar reference sphere (1737.4 km) and thus 
represents one of the most pronounced regional depressions on 
the Moon. In contrast, Mare Crisium exhibits voluminous 
basalt fillings. The terrain model indicates that the basin is 
slightly elliptical in shape, indicating, possibly, that the basin 
was formed by an oblique impact. Several large (>100 km- 
diameter) craters are visible in the terrain model. 
The Clementine terrain model of Mare Orientale extends from 
20°S-10°N to 88°W-90°W. It has a grid spacing of 200 m and 
a height resolution of better than 50m (Figs. 3, 4). Preliminary 
analysis of the terrain show a variety of surface features: 
depressions associated with lava ponds, isolated volcanic 
constructs, and steep slopes towards the center of Orientale up 
to 1500 m per 5000 m. Furthermore, a number of impact 
craters are included in the terrain model that show much 
morphologic detail for studies of depth, diameter, rim height 
and impact ejecta patterns. 
5. COMPARISON WITH LASER ALTIMETER DATA 
The Clementine laser altimeter obtained measurements of the 
slant range between the spacecraft and the lunar surface from 
spacecraft altitudes of 640 km or less (Nozette et al., 1994; 
Zuber et al., 1994). Laser pulses were emitted once every 0.6 
second. 19% of the shot returns were detected and resulted in a 
total of 114,000 altitude measurements. 
Altitude profiles along orbits are available, as well as an 
interpolated 2?x2? global raster terrain model. However, due to 
high spacecraft altitude, there is no data for lunar latitudes 
higher than 75?N or 75?S. For this study, the global data set 
was reprojected to match the stereo terrain model in terms of 
618 
map projection parameters and scale (Fig. 5). It is obvious that 
the resolution of the laser altimeter dataset is inferior to that 
of the terrain model from the stereo images. Furthermore, gaps 
in the dataset exist. Also, the laser altimeter data appears noisy 
in rugged terrain due to the difficulties in detecting reflected 
laser signals when these are scattered. In contrast, it is obvious 
that the stereo model is constrained by illumination and 
viewing conditions and becomes noisy near the lunar limb and 
the terminator. 
6. SUMMARY AND DISCUSSION 
The availability of digital image data and digital 
photogrammetric techniques has greatly improved our means to 
recover topographic information from spacecraft stereo image 
data. These datasets represent new and powerful tools for 
geologic studies of the large lunar impact basins. 
The study emphasizes that imaging sequences have to be 
planned carefully in terms of viewing geometry and lighting 
conditions if photogrammetric analyses are to be carried out. 
Orbit, camera pointing and imaging sequences have to be 
planned with the goal of achieving good image base-to-height 
ratios. However, lighting conditions are also important. 
Unlike Earth, albedo varies very little on the lunar surface, and 
there is little texture in images that were obtained at high Sun 
angles, therefore. Hence, this suggests that imagery should be 
obtained at low Sun angles, so long as cast shadows are 
avoided. Cast shadows pose a particular problem on lunar 
imagery because of the lack of an atmosphere and little 
scattering of light. In addition, we experienced that it is 
important for possible stereo partner images to be obtained 
under similar lighting conditions, as otherwise the automatic 
matching will fail. In addition, either good control point-, 
good spacecraft trajectory and camera pointing data, or both 
are needed for photogrammetric analysis. 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996