 |
| Stardust spacecraft schematic Camera periscope shown at bottom. |
New
27 June 2003
Back to [ A/CC Stardust page | A/CC News ]
Stardust 5535 Annefrank Flyby
By Bill Allen, A/CC editor & links librarian
Above: 5535 Annefrank during the 2 November 2002 Stardust spacecraft flyby. The original frames are credited to NASA/JPL, and were retrieved from the NASA Planetary Data System Small Bodies Node Stardust Mission archive. The animation was prepared by Bill Allen for the Asteroid/Comet Connection, as explained below. Columbine, Inc., holder of any rights inherent to original work done on this animation, hereby releases this file (intact with title frame) under the same Reproduction Guidelines for Use of NASA Images as apply to the original individual images.
Right: The spacecraft schematic is an illustration modified from the Stardust Mission Press Kit, a 623Kb PDF file.
Early on 2 November 2002 UT, but actually late evening Friday the 1st at JPL, the NASA Stardust comet mission flew past Main Belt asteroid 5535 Annefrank. The objective was to practice for the mission's comet 81P/Wild 2 flyby in January 2004, and the flight past the asteroid was planned to be at a distance of about 3,200 km. This would avoid any hazard to the spacecraft from possible companion objects or dust, but was too far to obtain good imagery.
As it turned out, 5535 Annefrank was a little larger than expected, and the distance was slightly less than planned. The result was interesting enough for the mission to publish one enhanced image amid some media fanfare on the 4th. That image is now enshrined on the Stardust Web site's Asteroid 5535 Annefrank page, and is also available as a 159Kb TIFF from the JPL public information office. See A/CC's links for news coverage from that time.
The mission reported on November 15th that all flyby images had finished being downloaded, but no further imagery was made available to the public until the pictures were archived early this year at the NASA Planetary Data System (PDS) Small Bodies Node section for Stardust Mission Support at the University of Maryland. They are stored in this folder as .img files in the PDS's own PDS3 file format, which is not generally supported by common graphics programs or Web browsers [note 1]. To prepare the page you are reading, where these images are shown publicly for the first time on the Web, a utility was used to get the files into a format useful to Adobe Photoshop, in which the images were prepared for export to a GIF animation program [note 2].
The asteroid was only 34 pixels wide at most in the original images, so, while the detail brought out by enlarging in Photoshop is impressive (it is shown at 400% on this page), the result is only somewhat informative and not reliable. It would be wonderful to see what could be done with high-end image processing tools and techniques. Both the stereo image below and the animation above hint at how working from multiple low-resolution images can yield higher resolution results, as well as rotation and 3D shape information.
 |
Two of the images — frames #448 and #449 — were taken about six seconds apart, both at the closest approach of 3,078.8 km. (1,912 miles), so they make a good stereo pair. To view the image at left, relax your eyes, let them lose focus while looking "behind" the screen, and bring the two white dots together into one dot at center. Above that center dot you will see the asteroid in stereo (and there will be images and dots to each side, but ignore those). Focus by moving your head toward or away from the screen, rather than by using your eye muscles.
Being there
The flyby planning is detailed in Stardust Mission Plan Post Launch Supplement B, Wild-2 Encounter Test: Annefrank Flyby, which is archived as a 580Kb PDF dated 23 October 2002. The mission's nicely detailed November 4th report tells how the flyby went, as does a JPL news release the same day, which describes 5535 Annefrank as being newly measured at "about 8 kilometers (5 miles) in diameter."
The image archive data set description contains a flyby timeline by frame number. The animation at the top of this page consists of frames from image sequences #40 and #41, most but not all of which are archived and are used here. It begins at frame #410 at 04:42:02 UT, at 4,718.3 km. (2,930.1 miles) from the asteroid. This first sequence becomes over saturated, "but these images were being taken to test the autotracking rather than for scientific purposes, and autotrack works perfectly well with saturated images." In fact this sequence is best for studying one side of the asteroid, an area that may show 5535 Annefrank to be a "contact binary" or even trinary, with one or two large formations that look to the untrained eye like they could be independent and held to the larger asteroid by gravity, perhaps sitting in dust accumulated from their interaction. (This proposition is only speculation from an illustrator who has spent a lot of time with these pictures in various forms, and, as noted above, enlargement interpolation can be misleading.)
The second sequence begins at frame #446 at 04:50:01 and 3,080.8 km. (1,913.2 miles), with exposure time reduced to obtain images "of scientific as well as engineering use." The last image in the archive is frame #476, taken at 04:53:01 — eleven minutes after frame #410 — at a distance of 3,300.5 km. (2,049.6 miles). There were about six seconds between each frame, but the animation runs 50 times faster than that, with intervals adjusted where some frames are missing.
  |
Beyond redistributing the original narrow tonal range over the full 8-bit grayscale range and enlarging the images by 400%, there is no other image manipulation in the above animation and stereo pair. At right are two images — frames #420 and #449 — that have had further work. Their contrast was adjusted to help bring out detail, and they had sharpening filters applied first at 800% size and then again at 50% (net 400% size). The original images are inset at original scale with the only manipulation being tonal redistribution.
Even if detail is hard to discern in these pictures, the reward is in getting to "be there," to visit one more asteroid on a very short list.
 |
Footnotes
- Free graphics programs for Linux users that can view some PDS formats are Gimp, which also has a Windows beta version, and XV. The free Java-based WebWinds is available for PDS viewing on many platforms. The author can't vouch for using any of these tools for viewing PDS3 image files.
Some common image editing programs can read RAW files, such as Photoshop with its File/Open As/RAW Options dialog shown at right. The settings shown here won't work with all images in the Stardust Annefrank archive, but do work with the 1.0Mb files that are most visually interesting. [back] - The PDS image format is similar to FITS, widely used in astronomical photography, so I modified my own FITS viewer/manipulator (written in the Python 2.0 scripting language) to read the PDS files, to redistribute their limited grayscale tonal range over a full range of 256 values (a lossless process), and to output 8-bit RAW files that could be read into Photoshop and further manipulated there. In Photoshop 5.0 the images were enlarged by 400%, aligned, cropped, and output to a GIF animation program. [back]
Resources