Generating digital terrain models using lroc nac images

GENERATING DIGITAL TERRAIN MODELS USING LROC NAC IMAGES

T. Trana, M. R. Rosiekb, Ross A. Beyercd, S. Mattsone, E. Howington-Krausb, M.S. Robinsona, B. A. Archinalb, K. Edmundsonb, D.
Harbourb, E. Andersona, and the LROC Science Team
a
Arizona State University, School of Earth and Space Exploration, 1100 S Cady, Tempe AZ, 85287 –
(thanh.n.tran@asu.edu)
b
United States Geological Survey, Astrogeology Science Center, 2255 N Gemini Dr, Flagstaff AZ. 86001 –(mrosiek,
ahowington, barchinal, kedmundson)@usgs.gov
c
Carl Sagan Center at the SETI Institute; dNASA Ames Research Center, Mail Stop 245-3 (Bldg. N245), Moffett Field,
CA, USA (Ross.A.Beyer@nasa.gov);
e
University of Arizona, Lunar and Planetary Lab, Tucson, AZ 85721

Commission VI, WG VI/4

KEY WORDS: DTM, LROC, topography, Moon, mapping

ABSTRACT:

The Lunar Reconnaissance Orbiter Camera (LROC) consists of one Wide Angle Camera (WAC) for synoptic multispectral imaging
and two Narrow Angle Cameras (NAC) to provide high-resolution images (0.5 to 2.0 m pixel scale) of key targets. LROC was not
designed as a stereo system, but can obtain stereo pairs through images acquired from two orbits (with at least one off-nadir slew).
Off-nadir rolls interfere with the data collection of the other instruments, so during the nominal mission LROC slew opportunities
are limited to three per day.

This work describes a methodology of DTM generation from LROC stereo pairs and provides a preliminary error analysis of those
results. DTMs are important data products that can be used to analyze the terrain and surface of the Moon for scientific and
engineering purposes. As of 12 September 2010, we have processed 30 NAC stereo pairs to DTMs with absolute control to the
Lunar Orbiter Laser Altimeter (LOLA) dataset. For the high-resolution stereo images (~0.5 mpp) from the primary phase, the DTM
vertical precision error and the elevation fitting error to the LOLA data is expected to be less than 1 meter. For the lower resolution
stereo images (~1.5 mpp) from the commissioning phase, the vertical precision error and elevation fitting error is expected to be 3
meters. This does not include an estimate of absolute error at this time. This will be included when the final LOLA data is available.
There are six independent groups generating DTMs (ASU, DLR/TUB, UA, USGS, OSU, and Ames), and collaboration will result in
a detailed error analysis that will allow us to fully understand the capabilities of the DTMs made from LROC datasets.

1. INTRODUCTION 1.2 LOLA

1.1 LROC NAC LOLA is designed to assess the shape of the Moon by
measuring precisely the range from the spacecraft to the lunar
The Lunar Reconnaissance Orbiter is currently in operation surface, incorporating precision orbit determination of LRO and
around the Moon (Chin et al. 2007, Vondrak et al, 2010). Two referencing surface ranges to the Moon’s center of mass. LOLA
instruments on board the spacecraft enable the extraction of has 5 beams and operates at 28 Hz, with a nominal accuracy of
digital terrain models (DTMs): Lunar Reconnaissance Orbiter 10 cm. One of its primary objective is to produce a global
Camera (LROC) and Lunar Orbiter Laser Altimeter (LOLA.) geodetic grid for the Moon to which all other observations can
LROC consists of one Wide Angle Camera (WAC) for synoptic be precisely referenced (Smith et al. 2010, Zuber et al. 2010.)
multispectral imaging, and two Narrow Angle Cameras (NAC)
to provide high-resolution images (0.5 to 1.5 m pixel scale) of 1.3 DTM Collection
key targets (Robinson et al. 2010a, 2010b.) LROC was not
designed as a stereo system, but can obtain stereo pairs through The LROC team has representatives from six different groups
images acquired from two orbits (with at least one off-nadir using four different methods to create DTMs.
slew). Typically the two observations that form a geometric
stereo pair have different slew angles ranging from zero to 1. Arizona State University (ASU)
twenty degrees (Beyer et al. 2009). To obtain an accurate DTM, 2. German Aerospace Center (DLR), Institute of Planetary
the convergence angle between the two images should be more Research and Technical University Berlin (TUB)
than 12º for images with a 0.5 pixel scale (Cook et al. 1996). 3. NASA Ames Research Center (Ames)
Off-nadir rolls interfere with the data collection of the other 4. University of Arizona (UA)
instruments, so during the LRO Exploration Systems Mission 5. Ohio State University (OSU)
Directorate primary phase LROC slew opportunities are limited 6. United States Geological Survey (USGS)
to three per day on average (Robinson et al. 2010b).

A special joint symposium of ISPRS Technical Commission IV & AutoCarto
in conjunction with
ASPRS/CaGIS 2010 Fall Specialty Conference
November 15-19, 2010 Orlando, Florida

ASU, NASA Ames, UA, and USGS all use SOCET SET km. Each camera has an internal buffer of 256 MB; allowing for
(DeVenecia et al. 2007) for photogrammetric processing the image length of 52,224 lines or 26,112 m. The NAC images are
NAC images, while OSU uses Orbital Mapper and Leica sampled at 12 bits, and companded to 8 bits. Each camera has
Photogrammetry Suite 9.3. DLR/TUB uses photogrammetry its own optics, and they are aligned to overlap by ~135 pixels in
software developed in-house. In addition, NASA Ames is using the cross-track direction and are offset from each other by ~185
the NASA Ames Stereo Pipeline, also developed in-house. All pixels in the down-track direction (Robinson et al. 2010b).
of these groups have successfully processed LROC images to
high precision DTMs. Models have been made of 30locales, Stereo images are collected by acquiring images on two
including the Apollo 15, 16, and 17 landing sites, as they different orbits so the total parallax angle is greater than 12°.
contain useful landmarks for absolute positioning (Davies and On average the parallax angle is about 24°. The overlap
Colvin, 2000, Archinal et al. 2010, Oberst et al. 2010), are of between the two NAC_L and NAC_R images provides three or
operational and scientific interest as ground truth sites, and are four stereo models from which to collect elevation data. The
of general interest due to their historical importance (Beyer et number of models depends on whether the areas where the right
al. 2010). and left images overlap are parallel or intersect each other
(Figure 1). The amount of overlap and the actual footprint are
Analysis by six groups using four techniques on similar data affected by the topography and the orbit parameters (target
allows an important initial comparison of derived camera center point, latitude, and slew angle).
parameters and an assessment of LROC DTM quality. Deriving
DTMs of areas that include positioning landmarks (Archinal et 2.2 LOLA
al. 2010) allows us to tie together LRO and other lunar datasets,
and to assist the Lunar Mapping and Modelling Project LOLA is a pulse detection time-of-flight altimeter that
(LMMP) (Noble et al. 2009) in deriving DTMs and controlled incorporates a five-spot pattern to measure the precise distance
mosaics of the Constellation Program regions of interest to the lunar surface at 5 spots simultaneously, thus providing 5
(Gruener and Joosten, 2009). profiles across the lunar surface for each orbit. LOLA fires at a
fixed, 28-Hz rate, so that for a nominal 1600 m/s ground track
The objective of this paper is to (i) describe the methodology of speed there is one shot approximately every 57 m. At a nominal
DTM generation from LROC NAC stereo pairs based on the 50-km altitude, each spot within the five-spot pattern has a
method being used by ASU, UA, USGS, and NASA Ames and diameter of 5 m while each detector field of view has a diameter
(ii) to discuss preliminary error analysis on the results. of 20 m. The spots are 25 meters apart, and form a cross pattern
canted by 26 degrees counterclockwise to provide five adjacent
2. DATA SOURCES profiles (PDS Geoscience Node 2010, Smith et al. 2010, Zuber
et al. 2010.) The LOLA instrument boresight is aligned with the
2.1 LROC NAC LROC NAC cameras to enable altimetry data collection in the
overlap region between the NAC_L and NAC_R.

Figure 1. Stereo Models Figure 2 - LOLA spot pattern

The LROC NACs are linear pushbroom cameras built using the Tracking of LRO is currently within 10 m radial and 300 m
Kodak KLI-5001G line array. The line array is a 5064 element horizontal accuracy (Zuber et al. 2010.) By using Earth-based
CCD with 7-micron pixels. The two NAC cameras are laser ranging tracking and crossover analysis, the expected
designated NAC-Left (NAC_L) and NAC-Right (NAC_R) and accuracy of the LOLA data will be 1 m radial and 50 m
these names are reflected in the image filename with NAC_L horizontal.
images having an L in the filename and NAC_R images having
and R in the filename. Each camera is designed to provide 0.5 3. METHODOLOGY
m resolution panchromatic images covering a 2,500 m swath
To generate the DTMs, we use a combination of the USGS
width, for a combined coverage of 5,000 m, at an altitude of 50
Integrated Software for Imagers and Spectrometers (ISIS) (see
in conjunction with

http://isis.astrogeology.usgs.gov/) and SOCET SET® from available to the science community when error analysis and
BAE Systems. ISIS routines ingest the image files, perform a documentation is complete.
radiometric correction, and export to a format SOCET SET
accepts. The NAC files imported into SOCET SET are Level 1 4. ERROR ANALYSIS
radiometrically corrected images and a list of keywords of
relevant parameters, such as spacecraft coordinates, altitude, The quality of a DTM is measured as a combination of absolute
Euler angles, and ephemeris positions. accuracy (latitude, longitude, and elevation at each pixel) and
relative accuracy (relative relief and slopes internal to the
SOCET SET uses a generic pushbroom sensor model to relate DTM). The LOLA data will be used to define the geodetic
the image space to ground coordinates. Often times, there is a reference frame for the DTMs extracted from the images. After
bias error in the camera pointing, which we correct with multi- a final crossover analysis is performed on the LOLA data and it
sensor triangulation (MST), more commonly known as bundle is adjusted to be self consistent, the adjusted data can then serve
adjustment, to update the parameters (position, velocity, as the best possible absolute accuracy geodetic reference frame
pointing angles, etc.) to improve the registration between to which the DTMs can be referenced. This section will focus
overlapping images and between images and ground truth. MST on the relative accuracy analysis of the DTMs and how well the
performs an aero-triangulation using sensor position, sensor DTMs are matched (referenced) to the LOLA elevation values.
pointing, ground points, and image tie points. Ground points tie The match between the LOLA elevation values and DTMs will
a point or identifiable object in the image to a point on the improve as errors in the LOLA data are reduced.
ground, and tie points relate a point in the overlap regions of
two or more images. Selected parameters, such as the position, 4.1 Expected Vertical Precision
velocity, and pointing angles are adjusted so that the residual
RMS for all ground and tie point measurements is minimized. Based on the spacecraft orbit and camera geometry, the
When working with a pair of LROC NAC stereo images, the theoretical expected vertical precision can be calculated. During
RMS residuals are typically ~0.25 pixels and are rarely larger the nominal mission, LRO is in a 50 km circular orbit. During
than 0.4 pixels. When working with multiple sets of stereo image acquisition, LRO is either pointed nadir or rotated about
images of the same region, there may be points with larger the flight line (normal situation) to acquire a stereomate. In
residuals, but the overall solution is typically under 1 pixel. some cases, LRO can also be pitched forward to acquire stereo
images in the polar regions. The convergence angle is the total
Once MST completes with an acceptable residual RMS, the parallax angle between the two stereo pairs. Based on the
process of extracting DTMs can begin with NGATE (SOCET spacecraft geometry, we can calculate the expected vertical
SET – Next Generation Automatic Terrain Extraction). NGATE precision using base to height ratio, assuming the instantaneous
performs image correlation and edge matching for every single field of view (IFOV) or ground sample distance (GSD) of 0.5
pixel in the image to create a dense model. The result is then re- m, and feature match RMS error of 0.2 pixels (Cook et al.
sampled to the desired DTM resolution (meters/post) that can 1996.). Figure 3 shows the expected vertical precision for
be anywhere between 3 to 10 times the pixel scale of the image convergence angle between 5 and 40 degrees.
to minimize noise. For images with moderate Sun (35º-65º
incidence angle), results from NGATE require very little
editing. However, images with areas of instrument saturation or
low Sun (large shadowed regions), and areas of high Sun where
albedo dominate morphological features require intensive
editing and interpolating across areas of no ground data.

NGATE is not optimized to work with linear pushbroom
images. One way to increase the effectiveness is to perform a
pair-wise rectification on the images that will be used in the
DTM extraction. This process rotates the images so that the
epipolar lines are horizontal and scales the images to a common
pixel scale. The rectified images make stereo vision easier on
the eyes and are required for accurate generation of the DTM.
Another way to increase the effectiveness is to generate a
continuous rational polynomial sensor model for the images.
The advantage of this method is that if pair-wise rectification is Figure 3 - Expected precision
used then 3 or 4 sets of images would need to be generated
(NAC_L – NAC_L, NAC_R – NAC_R, NAC_L – NAC_R,
NAC_R – NAC_L) depending on how many stereo models are
formed by the LROC NAC images (Figure 1.) The continuous
rational polynomial sensor model images can be combined with
different images since they are generated independent of any
other image.

Our standard products include the DTM and orthorectified
images in ISIS cube format. In addition, a hillshade image,
color shaded relief image, slope map, and confidence map are
provided in GeoTIFF format. These products will be made
in conjunction with

4.2 Spacecraft Jitter 4.3.1 Method 1: Matching based on geomorphic features

One method is to use geomorphic features that can be identified
in the images and the LOLA profiles. Bottoms of craters, tops
of hills, and breaks-in-grade can be matched between the stereo
images and the LOLA profiles. This method was used for
absolute control at the Apollo 15 site. The initial comparison
between an unedited DTM and the LOLA profiles showed that
the difference between LOLA and DTM elevation values
ranged from -88 to 245 m. By eliminating errors below -35 m
and above 30 m, only 55 out of 2,199 LOLA points were
eliminated from orbit 1576. These outliers will have to be
investigated to determine if the error is in the DTM or the
LOLA observations. By eliminating these high-error points the
range in difference values has a minimum of -34.7 and a
Figure 4 - Jitter in the Giordano Bruno DTM.
maximum of 25.6 m.
Spacecraft induced jitter during stereo image acquisition results
Table 1 – preliminary error analysis – Method 1
in artificial topographic ripples in some NAC DTMs (Mattson
Orbit # Points Average Std Dev RMS
et al. 2010.) We see these artifacts in about 10% of stereo
Error (m) Error (m) Error (m)
models taken during the LRO commissioning phase. The LROC
3060 1,613 1.4 1.7 2.2
Team and the LRO project are working to determine the cause
of the jitter and to mitigate the problem in the mapping phase of 1229 1,722 2.9 1.6 3.4
the mission (Mattson et al. 2010), and so far no jitter has been 2792 1,762 -3.1 2.0 3.7
detected in the ESMD primary mission. The artificial ripples are 1577 2,212 -2.8 3.3 4.3
readily apparent in derived shaded relief products with 1924 2,447 3.6 3.7 5.1
illumination along the down-track direction. For the Giordano 2887 981 7.4 5.8 9.4
*
Bruno DTM (Figure 4), jitter is seen as consistent banding in 1576 2,144 8.7 4.2 9.7
the cross-track direction, with higher frequency bands occurring 2445 2,195 10.1 4.4 11.1
every ~200 meters and lower frequency bands occurring every 2793 2,441 14.0 1.1 14.0
~800 meters. ALL 17,517 4.9 6.6 8.3
* 55 points were eliminated from this orbit
Most images do not have jitter and result in excellent DTMs.
Shown in Figure 5 is a subset of the color hill shaded mosaic Our preliminary vertical accuracy analysis for a DTM of the
for Lichtenberg crater. The entire DTM is based on 8 LRO Apollo 15 site is summarized in Table 1. The Apollo 15 DTM
NAC image pairs. has a pixel scale of 1.5 m generated from images with a pixel
scale of 0.5 m. For the bundle adjustment, there are 9 pairs of
LROC NAC images (18 NAC_L and NAC_R images) and 18
geomorphic points were picked to tie the images to the LOLA
elevation points. The NAC DTM is on average within 5 m for
five of the nine LOLA profiles (9,756 points). Since this DTM
is part of a larger bundle block adjustment, the other DTMs will
need to be examined to determine if a commensurate error
applies to other DTMs. The RMS errors for all the orbits are
comparable to the RMS errors reported in Table 2.

After registering the LOLA profiles and the DTM, the two
datasets were differenced and the residual is plotted as the
cumulative percent of points in the profile (Figure 5). The
profiles were ordered from lowest to highest RMSE. Orbits
2792, 2793, and 3060 are relatively flat lines with small tails
and have the lowest standard deviations shown in Table 1. The
time difference between sequential orbits (1576 and 1577, and
2792 and 2793) is about 2 hours, and it’s interesting to note that
Figure 5 - Color hill shade of Lichtenberg crater the error between the LOLA and DTM elevation values for
sequential orbits do not differ from the other orbits. The plotted
4.3 Error between DTM and LOLA lines show that the DTM is on average lower than LOLA
elevations.
There are different methods for connecting the image location
with the elevation values in the LOLA profiles, so the error
analysis differs.

in conjunction with

orthogonal direction, but in other packages (i.e. MATLAB,
ENVI, etc.), one can visualize both the LOLA profiles and the
DTM in a 3-dimensional state. This vantage point allows users
to know exactly where to place the geomorphic points so the
profiles and the DTM line up more accurately. Additionally, we
use MATLAB to perform autocorrelation between the
placement of the LOLA profile within the DTM and determine
which placement has the lowest error. The algorithm that we
currently use is in its early stages of development, and it
considers simultaneously changing 6 parameters: displacement
in the longitude (X), latitude (Y), and elevation (Z) direction,
rotation about the X and Z-axis, and scaling in the Y direction.
The result of this script is the image location of the control
points and the corresponding latitude, longitude, and elevation.

Figure 6 - LOLA profile plot The advantage of using the MATLAB script is it removes
ambiguity associated with the locations (image sample and line)
The LOLA vs DTM difference was overplotted on a NAC DTM of the geomorphic points. However, this technique is more time
derived shaded relief map to reveal systematic errors (Figure 6). consuming because it requires the user to extract and edit the
The area north of 26° latitude is a plain about -2100 m in DTM of an area overlapping the profile of interest before they
elevation. South of 26° is the beginning of Mons Hadley, rising can be imported into MATLAB. Currently, the LOLA profiles
3300 m above the plain. Within the flat region in the plain, are inconsistent with each other, so our strategy is to tie the
most of the absolute errors are less than 10 m and the majority DTM as best as we can with the help of MATLAB to one
of the errors are below 2 m. The errors are systematic and profile, which we called the primary profile. This ensures that
correlated within each profile. There is not any tilt apparent in the model is fixed in the down-track direction (Figure 7). The
the DTM relative to the LOLA points. This correlation shows other profiles (secondary profiles) are then used to tie the
that the DTM has been leveled with respect to the LOLA data. elevation of the model to control the possible tilt about the
In the Mons Hadley region the absolute errors have more primary profile (Figure 8). If the secondary profiles have a
variation indicating the possibility of a spatial offset between spatial and elevation offset from the primary profile, then the
the LOLA profiles and the DTM. After the LOLA profiles are error in the slope of the DTM could be up to 1° in the cross-
corrected with the crossover analysis, these errors should be track direction.
reduced. The DTM was not edited so some blunders in the
DTM account for some of the large errors. Table 2 -- Preliminary Error Analysis – Method 2
Locations Col. A Col. B Col. C
Aristarchus Plateau 1 (2 mpp*) 0.42 0.84 14.36
Gruithuisen Domes (2 mpp) 0.77 0.91 --
Ina D-Caldera (2 mpp) 0.43 1.20 10.46
Lichtenberg Crater (2 mpp) 0.48 1.04 5.23
South-Pole Aitken (2 mpp) 1.19 1.41 19.63
Hortensius Domes (5 mpp) 2.42 2.18 9.46
Marius Hills (5 mpp) 5.70 2.27 8.33
Reiner Gamma (5 mpp) 4.98 3.56 7.66
Sulpicius Gallus (5 mpp) 3.08 3.06 10.22
* meters per post (mpp)

The 2mpp DTMs are from high-resolution nominal phase stereo
images, and the 5mpp DTMs are from lower resolution
commissioning phase images (Table 2). The errors from
commissioning phase DTMs are higher than the errors from
nominal phase DTMs due to the limits of resolution.
A. Column A shows the precision error (m) reported in the
DTM header file generated by SOCET SET.
B. Column B shows the RMS error (m) between the primary
Figure 6 - Difference between LOLA and DTM LOLA profile and the DTM. This error provides insight on
how well the DTM and LOLA data agree.
4.3.2 Method 2: Matching based on aligning a profile C. Column C shows the RMS error (m) between all of the
LOLA profiles that overlap the DTM and the DTM.
Another method to match the DTM to the LOLA profiles is to Because the LOLA profiles have not been adjusted using a
use software capable of 3D plotting to allow visual rotation of crossover analysis, this error is dominated by the spatial
the DTM and LOLA data. The current version of SOCET SET and elevation offset between different LOLA data profiles.
has a powerful graphical interface that allows users to visualize
the depth of the stereo model, however, visualizing the depth
only gives a qualitative perspective and not a quantitative
perspective needed for precise placement of the LOLA profiles.
Within SOCET SET, one can only display images from the
in conjunction with

uses geomorphic points from one primary profile and elevation
control points from all other profiles. The slope error in the
down-track direction should be close to 0° because the RMS
error along the primary track is typically less than 1 m.
However, because of potential spatial and elevation offset from
secondary profiles, the slope in the cross-track direction can
have an error of up to 1°.

6. FUTURE WORK

NAC derived DTMs, along with the data from other sensors
onboard LRO, will be an important contribution to science
analysis. Therefore, it is important that the DTMs portray
Figure 7. Main profile from Aristarchus region (longitude ~- terrain as detailed and accurate as possible. Several ideas that
48.68°). The RMS error is 0.75 meters. we plan to implement include: (i) compare the DTMs of one
site made using two different datasets, (ii) compare the DTMs
that we made with DTMs that other groups made whether they
use the same technique or different technique, and (iii) compare
the DTMs with DTMs derived from Apollo era photographs.
The successful completion of these tasks will allow us to fully
understand the capabilities of the DTMs made from LROC
datasets.

For the LMMP mapping tasks, the DTMs that are generated
will be adjusted to be the final LOLA data after the LOLA data
is adjusted with a crossover analysis. This work is expected to
be completed in September 2011.

7. ACKNOWLEDGEMENTS
Figure 8. Secondary profile from Aristarchus region (longitude
~-48.81°). The RMS error is 18.94 meters. The high magnitude Thanks to the LROC and LOLA Science Operations Center
of the error reflects the spatial offset between the LOLA data Team, and the LRO Mission Management Team for all the hard
and the DTM most likely due to uncertainties in the current work and dedication to acquire the stereo images and altimetry
LRO ephemeris. data. Thanks to LMMP for support of map making.

5. PRELIMINARY ERROR ANALYSIS SUMMARY 8. REFERENCES

The absolute horizontal and vertical accuracy of the DTM Archinal, B.A., Duxbury, T.C., Scholten, F., Oberst, J., Danton,
largely depends on the absolute horizontal and vertical accuracy J., Robinson, M.S., Smith, D.E., Neumann, G.A., Zuber, M.,
of the LOLA data. The relative horizontal accuracy (pixel to LROC Team, LOLA Team, 2010. Tying LRO Data to the
pixel across the DTM) is the same as the spatial resolution of Fundamental Lunar Laser Ranging Reference Frame. 41st
the DTM. The absolute horizontal accuracy is the absolute Lunar and Planetary Science Conference, Houston TX, March
horizontal accuracy of the LOLA data. As of 12 September 2010, Abs. #2609.
2010, the absolute horizontal accuracy of the LOLA data can be
Beyer, R.A., Archinal, B., Li, R., Mattson, S., McEwen, A.,
up to 300 meters. It is expected that the absolute horizontal
Robinson, M.S., 2009. LROC Stereo Observations. LRO
accuracy of the LOLA data be as low as 50 meters with the
Science Targeting Meeting, Tempe AZ 2009, LPI Cont. 1483,
crossover analysis.
pp.15-16.
The vertical accuracy of the DTM is limited by both the Beyer, R.A., Archinal, B. Chen, Y., Edmundson, K., Harbour,
absolute vertical accuracy of the LOLA data and the Expected D., Howington-Kraus, E., Li, R., McEwen, A., Mattson, S.,
Vertical Precision (the relative precision) of the DTM. The Moratto, Z., Oberst, J., Rosiek, M., Scholten, F., Tran, T.,
absolute vertical accuracy of the LOLA data is expected to Robinson, M., LROC Team, 2010. LROC Stereo Data —
approach 1 meter, but as of 12 September 2010, the vertical Results of Initial Analysis. 41st Lunar and Planetary Science
accuracy of the LOLA is approximately 10 meters. The relative Conference, Houston TX, March 2010, Abs. #2678.
precision of the DTM from nominal phase is expected to be 0.5
meters, but can be as large as 1.5 meters. The relative precision Chin, G., Brylow, S., Foote, M., Garvin, J., Kasper, J., Keller,
of the DTM from commissioning phase is expected to be 3 m. J., Litvak, M., Mitrofanov, I., Paige, D., Raney, K., et al., 2007.
Lunar Reconnaissance Orbiter Overview: The Instrument Suite
Each technique used to match the DTM to the LOLA dataset and Mission. Space Sci Rev (2007) 129:391–419, DOI:
has merit. The first technique uses geomorphic points from all 10.1007/s11214-007-9153-y
profiles, so the final bundle adjustment solution has least square
error from all profiles. Because of discrepancies between the Cook, A.C., Oberst, J., Roatsch, T., Jaumann, R., Acton, C.,
LOLA profiles, there is a possibility of a slope error in both the 1996. Clementine Imagery: Selenographic Coverage for
down-track and cross-track direction. The second technique Cartographic and Scientific use. Planetary and Space Science,
in conjunction with

Volume 44, Issue 10, October 1996, Pages 1135-1148, ISSN Zagwodzki, T.W., 2010. The Lunar Orbiter Laser Altimeter
0032-0633, DOI: 10.1016/S0032-0633(96)00061-X. Investigation on the Lunar Reconnaissance Orbiter Mission.
Space Sci Rev (2010) 150: 209–241 DOI 10.1007/s11214-009-
Davies, M.E., Colvin, T. R., 2000. Lunar coordinates in the 9512-y
regions of the Apollo Landers. Journal of Geophysical
Research, Volume 105, Issue E8, 20,277–20,280, DOI: Vondrak, R., Keller, J., Chin, G., Garvin, J., 2010. Lunar
10.1029/1999JE001165 Reconnaissance Orbiter (LRO): Observations for Lunar
Exploration and Science. Space Sci Rev (2010) 150: 7–22 DOI
DeVenecia, K., Walker, A.S., Zhang, B., 2007. New approaches 10.1007/s11214-010-9631-5
to generating and processing high resolution elevation data with
imagery. Photogrammetric Week 2007, edited by D. Fritsch, pp. Zuber, M.T., Smith, D.E., Zellar, R.S., Neumann, G.A., Sun,
297–308, Wichmann, Heidelberg. X., Katz, R.B., Kleyner, I., Matuszeski, A., McGarry, J.F., Ott,
M.N., Ramos-Izquierdo, L. Rowlands, D., Torrence, M.H.,
Gruener, J.E., Joosten, B.K., 2009. NASA Constellation Zagwodzki, T.W., 2010. The Lunar Reconnaissance Orbiter
Program Office Regions of Interest on the Moon: A Laser Ranging Investigation. Space Sci Rev (2010) 150: 63–80
Representative Basis for Scientific Exploration, Resource DOI 10.1007/s11214-009-9511-z
Potential, and Mission Operations. Lunar Reconnaissance
Orbiter Science Targeting Meeting, Tempe, AZ June 9-11,
2009, Abstract #6036

Mattson, S., Robinson, M., McEwen, A., Bartels, A., Bowman-
Cisneros, E., Li, R., Lawver, J., Tran, T., Paris, K., LROC
Team, 2010. Early Assessment of Spacecraft Jitter in LROC-
NAC. 41st Lunar and Planetary Science Conference, Houston
TX, March 2010, Abs. #1871.

Noble, S.K., French, R.A., Nall, M.E., K. G., 2009. The Lunar
Mapping and Modeling Project. Houston TX Nov. 16-19, 2009
Annual Meeting of Lunar Exploration Analysis Group #2014

Oberst, J., Scholten, F., Matz, K.D., Roatsch, T., Wählisch, M.,
Haase, I., Gläser, P., Gwinner, K., Robinson, M.S., LROC
Team, 2010. Apollo 17 Landing Site Topography from LROC
NAC Stereo Data - First Analysis and Results. 41st Lunar and
Planetary Science Conference, Houston TX, March 2010, Abs.
#2051.

PDS Geosciences Node, 2010. LOLA RDR Query Page (Moon
ODE), Washington University in St. Louis,
http://ode.rsl.wustl.edu/mars/pagehelp/quickstartguide/lolardrqu
ery.htm (accessed 6 Sep. 2010)

Robinson M.S., Eliason, E.M., Hiesinger, H., Jolliff, B.L.,
McEwen, A.S., Malin, M.C., Ravine, M.A., Thomas, P.C.,
Turtle, E.P., Bowman-Cisneros, E., LROC Team, 2010a. Lunar
Reconnaissance Orbiter Camera: First Results. 41st Lunar and
Planetary Science Conference, Houston TX, March 2010, Abs.
#1874.

Robinson, M.S., Brylow, S.M., Tschimmel, M., Humm, D.,
Lawrence, S.J., Thomas, P.C., Denevi, B.W., Bowman-
Cisneros, E., Zerr, J., Ravine, M.A., Caplinger, M.A., Ghaemi,
F.T., Schaffner, J.A., Malin, M.C., Mahanti, P., Bartels, A.,
Anderson, J., Tran, T.N., Eliason, E.M., McEwen, A.S., Turtle,
E., Jolliff, B.L., Hiesinger, H., 2010b. Lunar Reconnaissance
Orbiter Camera (LROC) Instrument Overview. Space Sci Rev
(2010) 150: 81–124 DOI 10.1007/s11214-010-9634-2

Smith, D.E., Zuber, M., Jackson, G.B., Cavanaugh, J.F.,
Neumann, G.A., Riris, H., Sun, X., Zellar, R.S., Coltharp, C.,
Connelly, J., Katz, R.B., Kleyner, I., Liiva, P., Matuszeski, A.,
Mazarico, E.M., McGarry, J.F., Novo-Gradac, A., Ott, M.N.,
Peters, C., Ramos-Izquierdo, L.A., Ramsey, L., Rowlands,
D.D., Schmidt, S., Scott, V.S., Shaw, G.B., Smith, J.C.,
Swinski, J.P., Torrence, M.H., Unger, G., Yu, A.W.,

in conjunction with

Generating digital terrain models using lroc nac images

More Related Content

What's hot (20)

Similar to Generating digital terrain models using lroc nac images (20)

More from Sérgio Sacani (20)

Generating digital terrain models using lroc nac images