Mars Pathfinder

Science  05 Dec 1997:
Vol. 278, Issue 5344, pp. 1734-1742
DOI: 10.1126/science.278.5344.1734

The following foldout presents images and analysis from the Mars Pathfinder Mission that are discussed in seven subsequent Reports. The center is a four-page panorama of the surface of Mars around the lander (Plate 1). The back of the foldout contains surface images (Plate 7), a different perspective of the landing site (Plate 2), rover targets (Plate 3), locations of rocks and other features (Plate 6) and data analysis (Plates 4, 5, 8, 9, and 10).

Plate Captions

Plate 1.

Panoramic views of landing site from Sagan Memorial Station. Features are identified in Plate 6. Each view is a controlled mosaic of ~300 IMP images covering 360 of azimuth and elevations from ~4 above the horizon to 45 below it; the vertical coordinate is tan (elevation), equivalent to a central-perspective projection in one dimension, for optimum stereo viewing (B). Simultaneous least-squares adjustment of orientations of all images has been performed to minimize discontinuities between images. Mosaics have been highpass-filtered and contrast-enhanced to improve discrimination of details without distorting relative colors overall. Cartographic image processing by U.S. Geological Survey

A (Top) Enhanced true-color image created from the gallery pan sequence, acquired on sols 810 so that local solar time (LST) increases near-continuously from about 10:00 at the right edge to about 12:00 at the left. Mosaics of images obtained by the right camera through 670 nm, 530 nm, and 440 nm filters were used as red, green, and blue channels. The composite product was then slightly highpass filtered and contrast enhanced. Grid ticks indicate azimuth clockwise from north in 30 increments and elevation in 15 increments.

B (Bottom) Anaglyphic stereoimage created from the monster pan sequence, acquired in four sections between about 8:30 and 15:00 LST on sol 3. Mosaics of images obtained through the 670 nm filter (left camera) and 530 and 440 nm filters (right camera) were used where available. At top and bottom, left- and right-camera 670 nm images were used. Part of the northern horizon was not imaged because of the tilt of the lander. This plate may be viewed stereoscopically through glasses with a red filter for the left eye and a cyan filter for the right eye. To emphasize local topographic variations, parallax for points on the mean ground surface (but not elsewhere) has been removed during projection. The mean surface therefore appears to lie in the plane of the page, with topographic highs closer and lows beyond the page. [View Larger Version of this Image (89K GIF file)]

Plate 2.

Planimetric (overhead view) map of the landing site, to a distance of 20 m from the spacecraft. North is at the top in this and Plates 35. To produce this map, images were geometrically projected onto an assumed mean surface representing the ground. Features above the ground plane (primarily rocks) therefore appear displaced radially outward; the amount of distortion increases systematically with distance. The upper surfaces of the lander and rover also appear enlarged and displaced because of their height. Primary grid (white) is based on the Landing Site Cartographic (LSC) coordinate system, defined with X eastward, Y north, and Z up, and origin located at the mean ground surface immediately beneath the deployed position of the IMP camera gimbal center. Secondary ticks (cyan) are based on the Mars local level (LL) frame, which has X north, Y east, Z down, with origin in the center of the lander baseplate. Rover positions (including APXS measurements) are commonly reported in the LL frame. Yellow grid shows polar coordinates based on the LSC system. Cartographic image processing by U.S. Geological Survey. [View Larger Version of this Image (284K GIF file)]

Plate 3.

Overhead view of the area surrounding the Pathfinder lander illustrating the Sojourner traverse. Red rectangles are rover positions at the end of sols 130. Locations of soil mechanics experiments, wheel abrasion experiments, and APXS measurements are shown. The A numbers refer to APXS measurements as discussed in the paper by Rieder et al. (p. 1771). Coordinates are given in the LL frame.

The photorealistic, interactive, three-dimensional virtual reality (VR) terrain models were created from IMP images using a software package developed for Pathfinder by C. Stoker et al. as a participating science project. By matching features in the left and right camera, an automated machine vision algorithm produced dense range maps of the near field, which were projected into a three-dimensional model as a connected polygonal mesh. Distance and angle measurements can be made on features viewed in the model using a mouse-driven three-dimensional cursor and a point-and-click interface. The VR model also incorporates graphical representations of the lander and rover and the sequence and spatial locations at which rover data were taken. As the rover moved, graphical models of the rover were added for each position that could be uniquely determined using stereo images of the rover taken by the IMP. Images taken by the rover were projected into the model as two-dimensional billboards to show the proper perspective of these images. [View Larger Version of this Image (117K GIF file)]

Plate 4.

Topographic map of the landing site, to a distance of 60 m from the lander in the LSC coordinate system. The lander is shown schematically in the center; 2.5-m radius circle (black) centered on the camera was not mapped. Gentle relief [root mean square (rms) elevation variation 0.5 m; rms adirectional slope 4] and organization of topography into northwest and northeast-trending ridges about 20 m apart are apparent. Roughly 30 of the illustrated area is hidden from the camera behind these ridges. Contours (0.2 m interval) and color coding of elevations were generated from a digital terrain model, which was interpolated by kriging from approximately 700 measured points. Angular and parallax point coordinates were measured manually on a large (5 m length) anaglyphic uncontrolled mosaic (similar to Plate 1B but without least-squares adjustment of image orientations) and used to calculate Cartesian (LSC) coordinates. Errors in azimuth on the order of 1 are therefore likely; elevation errors were minimized by referencing elevations to the local horizon. The uncertainty in range measurements increases quadratically with range. Given a measurement error of 1/2 pixel, the expected precision in range is ~ 0.3 m at 10 m range, and ~10 m at 60 m range. Repeated measurements were made, compared, and edited for consistency to improve the range precision. Systematic errors undoubtedly remain and will be corrected in future maps compiled digitally from geometrically controlled images. Cartographic processing by U.S. Geological Survey. [View Larger Version of this Image (116K GIF file)]

Plate 5.

Mars-local-level (LL frame) coordinate map of rocks counted at the Mars Pathfinder landing site. Positions, apparent diameters (D), and heights (H) were measured to the nearest centimeter in the Marsmap virtual reality environment constructed from the Monster Pan set of IMP stereoimages. D is a perpendicular to a radial from the LL origin. All rocks with D 3 3 cm within a 3 m to 6 m annulus around the lander (total 1472) were used in the analyses. Rocks are assumed to be circular for coverage statistics estimates. Other rock counts were also made (DLR and Malin counts described in Smith et al., p.1758). [View Larger Version of this Image (85K GIF file)]

Plate 6.

Panoramic mosaic of the Mars Pathfinder landing site (compare with Plate 1A). Informal names of rocks and other features referred to in the Reports in this issue are indicated. Grid spacing is 30 in azimuth and 15 in elevation; vertical axis is linear in tan(elevation). Cartographic image processing by the U.S. Geological Survey and the Jet Propulsion Laboratory. [View Larger Version of this Image (53K GIF file)]

Plate 7.

Type areas of rocks and soils. (A) Dark rock type and bright soil type: Shown is the dark rock Barnacle Bill. Reflectance spectra typical of fresh basalt and APXS spectra indicating more silica-rich basaltic andesite compositions characterize this type. These rocks are typically the small boulders and intermediate-sized cobbles at the Pathfinder site. The bright soil type is very common and in this case comprises Barnacle Bill's wind tail and much of the surround soil area. This soil has a high reflectance and a strongly reddened spectrum indicative of oxidized ferric minerals. (B) Bright rock type: Shown is the bright rock Wedge. Reflectance spectra typical of weathered basalt and APXS spectra indicating basaltic compositions characterize this type. These rocks are typically larger than 1 m in diameter and many display morphologies indicating flood deposition. (C) Pink rock type: Shown is the pink rock Scooby Doo. APXS and reflectance spectra indicate a composition and optical characteristics similar to the drift soil. However, the morphology of the pink rock type indicates a cemented or rocklike structure. This material may be a chemically cemented hardpan that underlies much of the Pathfinder site. (D) Dark soil type: The dark soil type is typically found on the windward sides of rocks or in rock-free areas like Photometry Flats (shown here) where the bright soil has been striped away by aeolian action or in open areas. Other locations include the Mermaid Dune. (E) Disturbed soil type: The darkening of disturbed soil relative to its parent material, bright soil, as a result of changes in soil texture and compaction caused by movement of the rover and retraction of the lander airbag. (F) Lamb-like soil type: This soil type shows reflectance and spectral characteristics intermediate between the bright and dark soils. Its distinguishing feature is a weak spectral absorption near 900 nm not seen in either the bright or dark soils. [View Larger Version of this Image (151K GIF file)]

Plate 8.

Comparison of number and cumulative number of rocks versus rock diameter for MPF and the Viking landing sites (VL1 and VL2). The number of rocks per m2 for MPF is binned in 2D (m). Numbers indicate bin centers. The 1.414 m bin contains only Yogi. Cumulative number of rocks were calculated on a rock-by-rock basis and plotted at 1 cm intervals. VL diameters are the average of rock width and length. VL1 data do not include outcrops and neither VL line includes corrections for small rocks. Yogi produces a kink in the MPF line. Measurements in the far field using vertical stereo from pre- and post-deployment IMP images suggest an abundance of Yogi-sized rocks around 3104 m-2. [View Larger Version of this Image (28K GIF file)]

Plate 9.

Cumulative fraction of area covered by rocks with diameters 3 D within the annulus. Data are the same as for Plate 8. 16.1 of area is covered by rocks with D 3 3 cm for MPF. VL1 coverage is 5.6 and VL2 is 14.1. Rock area coverage in particular directions at MPF is highly variable (Plate 5); between azimuths 20 and 140, coverage is 10.2, and between 165 and 285 (in the Rock Garden) it is 24.6. Yogi contributes 1.7 to the area. In the far-field, Yogi-sized rocks probably cover 0.015 of the area. [View Larger Version of this Image (28K GIF file)]

Plate 10.

Cumulative area covered by rocks with heights 3 H around MPF and the VL sites. Data are the same as Plate 5. Because of the resolution in H, rocks with relief below about 1 cm were assigned H = 0.5 cm. [View Larger Version of this Image (28K GIF file)]


Navigate This Article