MESSENGER

Figure 1: Locations of MESSENGER image mosaics (Table 1) (outlined in white) and MESSENGER digital terrain models (brown shading) overlaid on a global MDIS image mosaic in cylindrical projection.

Introduction

The MErcury Surface, Space ENvironment, GEochemistry, and Ranging (MESSENGER) spacecraft is only the second probe to visit the innermost planet. The spacecraft is equipped with a well-calibrated imaging system (Hawkins et al., 2007, 2009), and data obtained during three Mercury flybys in 2008-09 (Solomon et al., 2008) included images that allowed stereo topographic reconstructions for a substantial portion of the planet not covered by Mariner 10 (Cook et al., 2000). The stereo analysis of the flyby images constitutes an important test case for MESSENGER's orbital mission phase, started in March 2011, when dedicated stereo mapping sequences will be obtained at near-global coverage.

MESSENGER's Mercury Dual Imaging System (MDIS) consists of two framing cameras, a wide-angle camera (WAC) and a narrow-angle camera (NAC), co-aligned on a pivot platform and equipped with identical 1024x1024-pixel charge-coupled device (CCD) sensors (Hawkins et al., 2009). The pivot-based scanning cap ability of MDIS allowed acquisition of several contiguous image mosaics (Fig. 1) during MESSENGER's three Mercury flybys on 14 January 2008 (M1), 6 October 2008 (M2), and 29 September 2009 (M3). These mosaics were constructed from a total of 2163 images, approximately 660 of which had a mean image resolution better than 550 m. The images were assembled into 10 individual sub-mosaics (Table 1). Five full-hemisphere sub-mosaics, M1 approach, M1 departure, M2 approach, M2 departure, and M3 approach, cover ~80% of Mercury's surface. In addition, MESSENGER acquired images from which five high-resolution local image mosaics were constructed: M1 H1, M1 H2, M2 H1, M2 H2, and M2 WAC H1 (Fig. 1).

Table 1: Overview of digital terrain models.
DTM Image mosaics Image scale (m) Image count Object point count (106) 3D point precision(m)
- M1 approach 520-580 38 - -
M1 DTM M1 H1 120-180 68 150.0 250.0
M1 H2 300-400 93
M1 departure 500-600 47
M2 DTM M2 WAC H1 250-750 5 220.0 290
M2 H1 100-300 35
M2 H2 250-350 173
M2 departure 500-650 47
M3 DTM M2 approach 500-550 20 34.5 160
M3 approach 450-500 28

From the images collected during the three flybys, there are three separate area covered stereoscopically (Fig. 1) by a total of 516 images (Table 1). The stereo mosaics were taken under similar illumination conditions but variable viewing conditions. Apart from images viewed nearly at nadir, large areas in these mosaics were located near the planetary limbs, where emission angles (measured from the local vertical) were as great as 80°. Stereo angle is an even more important parameter than emission angle for the generation of a high-fidelity digital terrain model (DTM). Stereo angles were appropriate for the M1 H1, M2 H1, M2 WAC H1, and M3 DTMs (up to 40° and 20-30°, respectively), but for the M1 H2 and M2 H2 mosaics stereo angles were often less than 14°, and the stereo angle was only 4° in the southwestern part of M1 H2 and M2 H2.

Figure 2: M1 DTM (hill-shaded, color-coded). Heights are given with respect to a sphere of

Method

The photogrammetric stereo analysis is based on algorithms and software realizations mainly developed for Mars Express and used extensively on previous planetary image data sets (Giese et al., 2006; Gwinner et al., 2010, Scholten et al., 2011, Preusker et al., 2011). The processing involves several stages and includes pointing corrections made with photogrammetric block-adjustment techniques, multi-image matching, and the generation of DTMs and orthoimage mosaics.

Navigation data correction

The navigation data correction was carried out using bundle block adjustment techniques. This type of least-squares analysis produces solutions for spacecraft position and camera pointing for each image on the basis of large numbers of tie-point measurements. The 2D image coordinates of the tie points for each image constitute the input (observations) to the photogrammetric block adjustment whereas the 3D coordinates of the surface points (ground points) and the orientation parameters (pointing and position) of each image were considered as unknowns. The nominal position and pointing were used as approximate values for the orientation data, as well as additional observations. This allows constraining the position along the spacecraft trajectory against the expected lower pointing accuracy (Zhang et al., 1996).

Image matching

The images were pre-rectified on a reference sphere with a radius of 2440 km using the improved orientation data, as described above. A multi-image matching technique was applied to derive tie points between images that form stereo observations. The pre-rectification warrants that the search for tie points be limited to small areas. Hence, point misidentifications and gaps were reduced to a minimum. The matching algorithm is an area-based image correlation to derive approximate values for the match-point coordinates, which are refined to sub-pixel accuracy by least-squares matching (Wewel, 1996). After the matching, the derived image coordinates (which refer to pre-rectified images) were transformed back to the geometry of the raw images, using the history files generated during the pre-rectification. The accuracy of this back-transformation is better than 0.1 pixel (Scholten et al., 2005).

Figure 3: M2 DTM (hill-shaded, color-coded). Heights are given with respect to a sphere of

DTM generation

Beginning with the large numbers of coordinate pairs for the matched points, the geometric calibration and improved orientation data were used to compute object point coordinates by means of forward ray intersection. Here, least-squares adjustment was applied for this over-determined problem. As a result, we obtained object point coordinates and their relative accuracy in Mercury body-fixed Cartesian coordinates. For the generation of a gridded DTM, the object points from the different stereo models were merged. The object points were first transformed from Mercury body-fixed Cartesian coordinates to geographic latitude/longitude/height and then transformed to chosen map projections (simple cylindrical equidistant). A pixel scale of 1 km was chosen. Object points located within a DTM pixel were averaged using neighborhood statistics (Gwinner et al., 2010). For regions that lack any object-point information, a gap-filling algorithm using DTM pyramids with reduced resolution was applied.

True orthoimage mosaics

Images were resampled to derive orthoimage mosaics. From the topography models, each image pixel was referenced to latitude and longitude by using ray intersection points with the terrain model. These true orthoimages are thus free of parallax errors and suited for the production of geometrically correct image mosaics.

Figure 4: M3 DTM (hill-shaded, color-coded). Heights are given with respect to a sphere of

Results

M1 DTM

The M1 DTM was derived from 208 stereo images acquired during the first flyby.  The images were combined into three individual sub-mosaics (Table 1). In total 241
individual matching runs were carried out on at least double- or triple-overlapping
images to yield 150 million object points with a mean intersection error of ± 250 m. The M1 DTM covers 12% (8.8·106 km²) of Mercury's surface (Fig. 2) and includes the Caloris impact basin (1550 km diameter). The coverage of the model is increased toward the limb over what was reported previously (Oberst et al., 2010) . Large portions of the DTM show topographic fabric consisting of relatively narrow, positive- and negative-relief landforms oriented radial to Caloris, most prominently expressed to the southwest and east-northeast of the basin (see arrows on figure). In addition, the DTM features a large and complex fault system near Beagle Rupes, one of the largest lobate scarps seen on Mercury (Watters et al., 2009a).

M2 DTM

The M2 DTM, the largest among the three DTMs, was derived from 260 stereo images acquired during the second flyby and includes four sub-mosaics (Table 1, Fig. 3). 220 million object points with a mean intersection error of ± 290 m were computed from 226 individual matching runs. The M2 DTM covers 15% (11.3·106 km²) of Mercury's surface and is limited by the positions of the limb and the terminator (at the time of M2) to the east and west, respectively. The DTM covers mostly heavily cratered terrain, in contrast to the M1 DTM. A large (~800-km-diameter) unnamed impact basin (centered at 16.4° N, 19.6° E), not evident in the corresponding images because of low incidence angles (measured from the local vertical), is a prominent feature in the DTM. Several more highly degraded basins are visible in the DTM. Like the M1 DTM, this second DTM also shows a number of high-relief fault structures, some extending over distances of up to several hundred kilometers (Fig. 3).

M3 DTM

Finally, the M3 DTM was produced from a combination of the two approach mosaics
constructed from images acquired during the second and third flybys. Owing to favorable stereo geometry with stereo angles of ~25°, the geometric accuracy of this DTM is the best among the three. However, because the two mosaics were taken at different local times, i.e., different illuminations, the stereo matching is subject to error in regions with shadows. In total 48 stereo images were used to compute 34.5 million object points with a mean intersection error as small as ± 160 m. This smallest among the three DTMs is elongated, extending from high northern to high southern latitudes, and covers ~5% (4.5·106 km²) of Mercury's surface. Prominent in this DTM is the Rembrandt impact basin, approximately half of which is covered by the terrain model (Fig. 4). The DTM also includes several prominent lobate scarps in the southern hemisphere, including the longest scarp yet found on Mercury that crosscuts the Rembrandt basin (Watters et al., 2009b).

General attributes

All DTMs were produced with a common grid spacing of 1 km, a value chosen to yield ~15 object points per DTM pixel on average. The three separate DTMs were merged to a single global DTM (Fig. 5), which covers approximately one third of Mercury's surface. Although the DTMs were generated separately, average heights and  topographic trends appear consistent. The total range in heights over all DTMs is approximately 9.5 km (-4.5 km to 5 km relative to the planetary datum). The average DTM height is 260 m, implying an average planetary radius for the regions of these DTMs of 2440.3 km. The terrain models feature large numbers of impact craters that
span a range of sizes and degradation states. We estimate that a total of ~400 craters larger than 15 km in diameter are included in these three models.

Figure 5: Global representation of all three flyby DTMs. These DTMs cover ~30 percent of Mercury's surface.

Conclusion

The stereo DTMs produced from the MESSENGER flybys will allow us to carry out a variety of morphologic studies of surface features on Mercury. The DTMs, for instance, include 400 craters larger in diameter than the effective resolution limit of 15 km, which will provide a basis for in-depth morphological study.
During its orbital mission phase, MESSENGER will obtain dedicated stereo observations under ideal Illumination and viewing condition (incidence angles between 5°and 75° and low emission angles), and optimized parallax angles (approximately 20°) for global topographic models of reduced noise level and higher spatial resolution. Also, we expect that with global coverage the image blocks will be more stable and able to overcome residual height offsets and long-wavelength trend errors in the terrain models. The currently available models constitute important tools for a variety of geological studies and will provide new insight into Mercury's surface morphology and tectonics.

Acknowledgements

The MESSENGER project is supported by the NASA Discovery Program under contracts NASW-00002 to the Carnegie Institution of Washington and NAS5-97271 to the Johns Hopkins University Applied Physics Laboratory.

Downloads

Table 2: Downloads

Data set: DTM pds format Data set: DTM ISIS3 fotmat Updated
M1.EQUI180_1km.float.dtm.pds M1_DTM.cub 5th Oct 2011
M2.EQUI000_1km.float.dtm.pds M2_DTM.cub 5th Oct 2011
M3.EQUI000_1km.float.dtm.pds M3_DTM.cub 5th Oct 2011

References

  • Cook, A.C. et al., 2000. Mariner 10 stereo image coverage of Mercury. J. Geophys. Res. 105, 9429-9443.
  • Giese, B. et al., 2006. Topographic modeling of Phoebe using Cassini images. Planet. Space Sci. 54, 1156-1166.
  • Gwinner, K. et al., 2010. Topography of Mars from global mapping by HRSC high-resolution digital terrain models and orthoimages: characteristics and performance. Earth Planet. Sci. Lett. 294, 506-519.
  • Hawkins, S.E., III et al., 2007. The Mercury Dual Imaging System on the MESSENGER spacecraft. Space Sci. Rev. 131, 247-338.
  • Hawkins, S.E., III et al., 2009. In-flight performance of MESSENGER's Mercury Dual Imaging System. In: Hoover, R.B., Levin, G.V., Rozanov, A.Y., Retherford, K.D. (Eds.), Instruments and Methods for Astrobiology and Planetary Missions. SPIE Proceedings, vol. 7441, paper 7441A-3, 12 pp., SPIE, Bellingham, Wash.
  • Oberst, J. et al., 2010. The morphology of Mercury's Caloris basin as seen in MESSENGER stereo topographic models. Icarus 209, 230-238.
  • Scholten, F. et al., 2005. Mars Express HRSC data processing - methods and operational aspects. Photogramm. Eng. Remote Sens. 71, 1143-1152.
  • Scholten, F. et al., 2011. GLD100 -The Global Lunar 100 Meter Raster DTM From LROC WAC Stereo Models, LPSC 2011, #2046.
  • Solomon, S.C. et al., 2008. Return to Mercury: a global perspective on MESSENGER's first Mercury flyby. Science 321, 59-62.
  • Watters, T.R. et al., 2009a. The tectonics of Mercury: the view after MESSENGER's first flyby. Earth Planet. Sci. Lett. 285, 283-296.
  • Watters, T.R. et al., 2009b. Evolution of the Rembrandt impact basin on Mercury. Science 324, 618-621.
  • Wewel, F., 1996. Determination of conjugate points of stereoscopic three line scanner data of Mars 96 mission. Int. Arch. Photogram. Remote Sensing 31, 936-939.
  • Zhang, W. et al., 1996. Methods of the bundle block adjustment of planetary image data. Int. Arch. Photogrammetry Remote Sens. 31, 1004-1009.

Last update: 06/10/2011 17:25