International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B4, 2012
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia
Comparison of Lunar Topographic Models Derived from Multiple Sources
Based on Least Squares Matching
J. Guo*, B. Wu
Department of Land Surveying & Geo-Informatics, The Hong Kong Polytechnic University, Hung Hom, Kowloon,
Hong Kong — jian.guo@polyu.edu.hk
Commission IV, WG IV/7
KEY WORDS: Lunar Topographic Model, Comparison, Surface Matching, Least Squares, Chang'E-1, SELENE
ABSTRACT:
Lunar topographic information is of paramount importance for lunar exploration missions and lunar scientific investigations. For
lunar topographic models derived from different sources or missions, it is not uncommon to see inconsistencies among them. This
paper presents a detailed comparative analysis of lunar topographic models derived from the Chang'E-1 and SELENE altimeter data.
A least squares matching method is developed to register one topographic model to the other through a seven parameters
transformation (three translation parameters, three rotation parameters, and one scale factor). Two typical study areas on the Moon
are selected for detailed investigation, including the Sinus Iridium (the primary candidate landing site area for future Chinese robotic
or human landed missions) and the Apollo 15 landing site area. Experimental results indicate that there are small shifts in both
horizontal and vertical directions between these two data sets. After the least squares matching and shift removal, the topography
derived from the two data sets show a consistent trend.
1. INTRODUCTION
Lunar topographic information is of paramount importance for
lunar exploration missions and lunar scientific investigations.
Starting from the 1960s, a vast amount of lunar topographic
data, including lunar images and laser altimeter data have been
collected and processed in the Apollo missions (Mellberg, 1997)
and the Clementine mission (Smith et al., 1997; Rosiek et al.,
1999). More recently, China launched its first lunar probe
Chang’E-1 in October 2007 (Ouyang et al., 2008; Zheng et al.,
2008). The Japanese lunar mission SELENE (SELenological
and ENgineering Explorer) was launched in September 2007
(Kato et al, 2008). The United States also successfully
launched its LRO (Lunar Reconnaissance Orbiter) to the Moon
in June 2009 (NASA, 2009). Among the payload on-board the
Chang’E-1, SELENE, and LRO, there are sensors (e.g.
cameras and laser altimeters) collect lunar topographic data at
various levels of resolution. These new data sets enable a new
era of lunar topographic modelling with the capabilities of
providing more detailed and precision lunar topographic
information.
For lunar topographic models generated from the imagery or
laser altimeter data from different sources or missions, it is not
uncommon to see inconsistencies among them due to
unavoidable reasons (e.g., different data collecting times, errors
from sensor positions and orientations, and different data
processing methods). Systematic study of the inconsistencies
between the lunar topographic models derived from different
sources will be helpful to evaluate the accuracies of the
topographic data and understand the performances of the
sensors. This paper presents a detailed comparative analysis of
lunar topographic models derived from different sources (1.e.,
from Chang'E-1 and SELENE altimeter data) based on least
squares matching method. The longitude, latitude, and altitude
coordinates in the Mean-Earth/polar axis (ME) coordinate
system of the laser altimeter points are used for analysis for
both the Chang’E-1 and SELENE data sets.
After giving a literature review on previous related works, a
least squares matching method employing seven transformation
parameters (three translation parameters, three rotation
parameters, and one scale factor) is developed for the
comparative analysis of different lunar topographic models. The
Chang’E-1 and SELENE laser altimeter data at the Sinus
Iridium area and Apollo 15 landing site are employed for
detailed experimental analysis. Finally, concluding remarks are
presented and discussed.
2. RELATED WORK
The topography of the Moon has been measured by several
approaches, including radar interferometry, sterco-
photogrammetry, and laser altimetry. Because of the Moon’s
synchronous rotation, most early lunar topography results were
restricted to the nearside. For the far-side of the Moon, both the
accuracy and resolution were tremendously poor. Smith et al.
(1997) reported the first reliable near-global topographic model
of the Moon, the GLTM2, which was obtained by the
Clementine LIDAR mission in 1994. USGS combined Earth-
based radar interferometry data, Clementine stereoimages
measurements, and Earth-based photos from Apollo, Mariner
10, and Galileo to generate the Unified Lunar Control Network
2005 (ULCN 2005), which is the latest global control network
of the Moon (Archinal, 2005). More recently, a mass amount of
lunar topographic data has been collected by the new lunar
missions in the past several years, e.g., the Chang’E-1,
SELENE, and LRO missions.
For the comparative analysis of lunar topographic models
derived from different sources, iz et al. (2011) examined the
consistency of the Chang’E-1 and SELENE reference frames In
the global scale through the analysis of a large number of nearly
co-located laser altimeter points from these two missions. A
twelve parameter affine transformation model was used for
comparison in the form of rigid body motions and
296
deft
mot
fran
tran
stra
reve
0.9»
fran
Pin
mea
proc
CL]
othe
data
has
Cha
didn
spar
com
Cler
6.8
app!
reso
resp
larg
diff
side
com
is ar
mod
mea:
Li €
genc
SEL
low
data
exan
displ
high
DEM
than
the |
The
high
For
topo,
past.
dime
proc
moti
(Iter:
basic
point
iterai
pairs
error
calcu
Othe
base
and
regis
meth
the n
