> XXXIX-B4, 2012
-Zoom mode. The PDS
evel 1 calibrated byt
2 map-projected (CDR-
o date. The volume of
other products contain
because of redundancy
zhly 3 orders of magni-
Mini-RF was added to
istration in 2005.
IS
vious section include
ll as geometrically raw
^ majority of these
r at best semicontrolled.
DTMs, and other map
among the controlled
taset, for which orbit
S have been adjusted to
where altimetric profiles
re dense gridded data
poles (Mazarico et al.
LA, and the ability of its
5 well as elevations, this
ence for other mapping
Iso being corrected by
ough so far on a local
ingly, studies exploring
of the altimetry from
ium et al. 2012) have
d using the altimetric
ntrolled image products
| of M? data based on
production of polar (Lee
(Rosiek et al. 2012)
iges. The polar mosaics
, — the largest controlled
covering the lunar polar
xel. Much of the latter
yA Lunar Mapping and
|. 2009; data accessible
nasa.gov/). A variety of
is such as orthoimages
Ms from Chandrayaan-1
Chang’e-1 CCD images
es (see Beyer et al. 2011
producing such DTMs
nsorship, and Rosiek et
lore recent summaries).
Chandrayaan-1 Mini-RF
(Kirk et al. 2011) and
is currently under way
RENT DATA
nd exploration, the lunar
st be co-registered in a
)nly such an effort will
yn, and error analysis of
ie full comparative and
ummary in this section
. the development and
a products.
| from all lunar missions
mon frame via geodetic
c, radargrammetric, and
s rigorous process will
ecessary to generate the
al DTM. Such a model
ic calibration and ortho-
nages are brought into
International Archives of the Photogrammetry, Remote Sensin
g and Spatial Information Sciences, Volume XXXIX-B4, 2012
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia
common frame and a common DTM is in use, the datasets can
be converted into information, primarily in the form of useful
cartographic products. Such products are essential for
addressing lunar science and exploration goals at the highest
possible level of accuracy. As a result of the merging process,
the accuracy level of such products will be known and
documented, which will be critical for the comparison of the
products and for their use in future decision making.
In order to meet the increasing needs of the science and
exploration communities, datasets must be comparable at the
pixel level with accuracy on the order of tenths of a pixel
required for color and spectral data. Such accuracy is only
possible with geodetically controlled products that are
orthorectified onto DTMs with resolutions approaching those of
the output image products.
Detailed arguments have been put forth that more extensive
cartographic efforts are needed to exploit past missions fully
and to prepare properly for future missions (Archinal et al.
2007; Kirk et al. 2008). The NASA Advisory Council has
recognized the importance of such processing, recommending
that all lunar datasets be geodetically controlled (NAC, 2007).
The IAU Working Group on Cartographic Coordinates and
Rotational Elements has also recently recognized the value of
controlled products (Archinal et al. 2011, recommendation 1)
and the need to generate them from new mission datasets. As
noted in Section 2, controlled cartographic products from recent
missions are greatly outnumbered by uncontrolled products.
The number of controlled products is growing and efforts to
combine data from multiple missions have begun (e.g., Iz et al.
2011; Shum et al. 2012) but given the volume and complexity
of the data it is clear that a massive effort will be required to
control even the most critical of these new large lunar datasets.
Given the funding constraints on recent major international
missions to the Moon and the need to register datasets from
multiple missions, an international co-operative project would
greatly facilitate accomplishment of the work described here. If
necessary, significant progress could be made even without
requiring the release of raw data from all missions. Joint efforts
at mapping would be a good first step that would greatly
encourage and facilitate broader international cooperation in the
exploration of the Moon.
In the following subsections we describe the need for
controlling the data, for generating a merged global DTM, and
for establishing a common reference frame. Basic high
resolution datasets are listed that need to be connected initially
and principles of processing are described to outline in what
order and how datasets could be registered to each other and a
common frame. Some of the many and difficult challenges in
accomplishing such work are briefly considered.
31 The Need for Geodetic Control
The only way to connect/register/compare data with quantified
precision and accuracy is to geodetically (usually photo-
grammetrically) process the data into controlled products.
Otherwise the uncertainties in the comparison of datasets
undermine their synergistic value. Users always want the best
precision and accuracy possible and require that they be
quantified. Such knowledge is critical for mineralogic, geologic,
and other scientific investigations and exploration purposes
Such as site selection, landing, and landed operations. Con-
tolling any single dataset provides many benefits including: (a)
the best method of removal of mosaic seams for qualitative
Work; (b) proper orthometric projection of data (i.e., registration
of Images to topography in order to make or match existing
mosaics and maps); (c) registration of multispectral data, which
Is essential to do at subpixel precision to avoid fringing
artifacts; and (d) proper photometric correction of data. The
Value of such control increases exponentially when multiple
datasets are considered, so it is essential that this work be
Planned for and done with new lunar data. Geodetic control
adds substantial value to the data, especially relative to the cost
of data collection and the immense risk that future surface
46
missions may fail if the maps used to evaluate landing site
safety or plan their operations are insufficiently accurate.
3.2 The Need for Global Topography
As noted in Section 2, new global DTMs have recently been
produced from Kaguya (Araki et al. 2009), LOLA (Smith et al.
2010), and Chang’e-1 (Li et al. 2010) altimetry, as well as
Kaguya TC (Haruyama et al. 2012) and LROC WAC (Scholten
et al. 2011) stereo imagery. As revolutionary and scientifically
valuable as these models are, there is still a need for global
topographic modeling at higher resolution and accuracy. For
example, the laser altimetry models have substantial longi-
tudinal data gaps at mid- and particularly equatorial latitudes.
The WAC stereo DTM is based on ~100 m resolution images
that, although aligned with LOLA Team derived spacecraft
position information, are uncontrolled and may have errors
comparable to their resolution. These existing global models are
therefore insufficient for the orthoprojection of high resolution
images at or even near the resolution of such data. They are also
insufficient for the orthoprojection, slope correction, and
calibration of medium resolution (100 m/pixel or more) color,
multispectral, or infrared data (e.g., Kaguya MI and SP, LRO
WAC and DLRE, Chandrayaan-I M’). Correction of slope
based photometric effects requires topographic data with
horizontal resolution at the image pixel scale or less and vertical
precision on the order of a tenth of a pixel or less. Such
photometric correction has been shown to affect the
compositional interpretation of spectral data at the 5% level and
significantly affect geologic interpretations of spectral
variability (Robinson and Jolliff 2002).
A high-resolution, global DTM is not only needed to process
global datasets in preparation for scientific analysis, it is critical
for successfully planning and conducting robotic and human
mission operations on the Moon. Even higher resolution DTMs
are needed to process local to regional high-resolution data.
Such DTMs can be generated from the combination of the
altimeter data and stereo data, particularly (in order from
highest to lowest resolution) NAC, Apollo, TMC, CCD-2, TC.
MI, and LRO Mini-RF imagery.
3.3 What System and Frame?
The recommended coordinate system for the Moon (Archinal et
al. 2011; LRO & LGCWG 2008) is the mean Earth / polar axis
(ME) system, and the recommended way to access it is via the
JPL DE 421 ephemerides, with an appropriate rotation to the
ME system. The recommended mean radius for the Moon is
1737.4 km (Archinal et al. 2011; LRO & LGCWG, 2008), and
fortunately most instrument teams and missions have adopted
these recommendations. The real issue then becomes using or
creating a reference frame within that coordinate system to
which datasets can be referred. Currently the best lunar
reference frames are those derived from Lunar Laser Ranging
(LLR). These frames have coordinate System accuracies
approaching the decimeter to centimeter level, but only for the 5
existing LLR targets. It will be necessary to tie the other
datasets into an LLR frame or one based on it.
3.4 What Datasets?
Noted above are some of the highest density or resolution
altimetric and stereo datasets that can be used to build a
fundamental lunar reference frame and uniform global DTM.
Other required data include spacecraft geometric (“SPICE” —
Acton 1999 —or similar) data and a lunar gravity model (ideally
incorporating the results from the GRAIL mission). Once such a
frame and model are established, all lunar data can be tied to
them, including the recent mission data described in Section 2,
and data from earlier missions such as Lunar Orbiter, Apollo,
Clementine, and Lunar Prospector.
3.5 Processing Principles
Some flexibility exists concerning the order in which data
should be processed, and in which algorithms, software, and
1
