This morning we have an image of MU69, the Kuiper Belt object to which New Horizons is heading, with arrival and flyby scheduled for January 1, 2019. This just after the first glimpse of the asteroid Bennu by the spacecraft now heading there for observation and sample return, OSIRIS-REx. By way of comparison, the first glimpse New Horizons had of Pluto/Charon came during an optical navigation test using the Long Range Reconnaissance Imager (LORRI), which occurred in September of 2006 when Pluto was still 4.2 billion kilometers away.
We knew in the year of its launch, in other words, that New Horizons could find and track targets at extremely long range, but MU69, otherwise known as Ultima Thule, is a tiny target indeed. Moreover, it’s one that raises a host of obstacles particularly in terms of the background stars. We are trying to pluck it out of field objects from a distance of 172 million kilometers.
“The image field is extremely rich with background stars, which makes it difficult to detect faint objects,” said Hal Weaver, New Horizons project scientist and LORRI principal investigator from the Johns Hopkins Applied Physics Laboratory in Laurel, Maryland. “It really is like finding a needle in a haystack. In these first images, Ultima appears only as a bump on the side of a background star that’s roughly 17 times brighter, but Ultima will be getting brighter – and easier to see – as the spacecraft gets closer.”
Image: The figure on the left is a composite image produced by adding 48 different exposures from the News Horizons Long Range Reconnaissance Imager (LORRI), each with an exposure time of 29.967 seconds, taken on Aug. 16, 2018. The predicted position of the Kuiper Belt object nicknamed Ultima Thule is at the center of the yellow box, and is indicated by the yellow crosshairs, just above and left of a nearby star that is approximately 17 times brighter than Ultima.
At right is a magnified view of the region in the yellow box, after subtraction of a background star field “template” taken by LORRI in September 2017 before it could detect the object itself. Ultima is clearly detected in this star-subtracted image and is very close to where scientists predicted, indicating to the team that New Horizons is being targeted in the right direction. The many artifacts in the star-subtracted image are caused either by small mis-registrations between the new LORRI images and the template, or by intrinsic brightness variations of the stars. At the time of these observations, Ultima Thule was 172 million kilometers (107 million miles) from the New Horizons spacecraft and 6.5 billion kilometers (4 billion miles) from the Sun. Credit: NASA/Johns Hopkins University Applied Physics Laboratory/Southwest Research Institute.
We are now in the process of continually setting records. These images are the most distant ever taken from the Sun, an honor that was once held by Voyager 1 in its famous ‘Pale Blue Dot’ image of the Earth, taken in 1990. New Horizons likewise snagged the award for most distant image from Earth in December of 2017. The images above were taken on August 16, a series of 48 that succeeded at the spacecraft’s first attempt to find MU69 with its own cameras. Now we can spend the next four months updating the record and watching MU69 grow.
And since ‘first glimpse’ images seem to define this week, here’s that first New Horizons image of Pluto.
Image: The Long Range Reconnaissance Imager (LORRI) on New Horizons acquired images of the Pluto field three days apart in late September 2006, in order to see Pluto’s motion against a dense background of stars. LORRI took three frames at 1-second exposures on both Sept. 21 and Sept. 24. Because it moved along its predicted path, Pluto was detected in all six images. The images appear pixelated because they were obtained in a mode that compensates for the drift in spacecraft pointing over long exposure times. Credit: NASA/Johns Hopkins University Applied Physics Laboratory/Southwest Research Institute.
This far away from home, it seems fitting to quote from the one Edgar Allen Poe verse, ‘Dreamland,’ that mentions Ultima Thule in its incarnation as a literary figure for the most distant of Earthly places:
By a route obscure and lonely,
Haunted by ill angels only,
Where an Eidolon, named Night,
On a black throne reigns upright,
I have reached these lands but newly
From an ultimate dim Thule –
From a wild weird clime, that lieth, sublime,
Out of Space – out of Time.
I am so excited about MU69. The Edgar Allen Poe quote is perfect–and we (or more precisley the NASA crew) are the angels haunting New Horizons. Hey, what’s the story on the length of those exposures: 29.967 seconds? There must be a reason for such a weird duration so close to 30s.
This document http://www.boulder.swri.edu/pkb/ssr/ssr-lorri.pdf , which seems pretty authoritative, says that LORRI exposure times are nominally 50-200 ms with 100 ms being the design spec for the instrument.
http://www.planetary.org/blogs/emily-lakdawalla/2018/0829-osiris-rex-and-new-horizons-first-light.html
There’s a link on this page that goes into how ‘they’ performed a software update to change LORRIs exposure times.
Thank you Mark Zambelli, that is very interesting! I am still curious why the maximum exposure length is now exactly 1/(the NTSC frame frequency) but let’s put it down to coincidence.
There are two opposite constraints on the exposure time. Too long exposures create streaks instead of points of light, due to the target object movement. Too short exposures can’t register faint objects. So for different points on NH trajectory you obtain different best exposure times. Probably is simply that, there is no special meaning in the number.
Thank you for the article, looking forward for a closer look on this very distant world. My interest is most about it’s surface composition, will there be organic tar or frozen gas?
Possibly “thulins” : ^ )
Great article.
Re Andrei’s question: I’m no astronomer, but 29.97 is the normal frame rate for US TV signals based on a standard called NTSC (it was originally 30 frames per second but reduced ever so slightly when colour signals were introduced).
https://en.wikipedia.org/wiki/NTSC
Thanks, your Wikipedia link allowed me to learn why that strange number is used in the original NTSC spec: the horizontal line frequency had to be decomposable into small integers: 9/572 Mhz (3^2 / 2^2.11.13), and the number of horizontal lines drawn each frame is an integer: 525.
But now I suspect that the NTSC frame rate has found its way into Paul’s article (or its source) by mistake: surely it’s much too big a coincidence for the exposures to have been 29.967 seconds long, when the NTSC frame rate is 29.967 Hz?
Just for clarification, the figure comes from JHU/APL and is in their caption of the imagery.
It is interesting that the subtracted imaged does not show an artifact for the star behind the target. The artifacts seem to visually match the stars in the raw image. I can only imagine that the registration was very carefully matched to the target and the star behind it, allowing for the best subtraction to eliminate it.
I would like an expert at this to explain the true reason for this effect though.
I think that is due to the algorithm being adjusted to eliminate that particular star. For example, if all stars are a little offset from their original positions but by different amounts/directions, you can slightly “move the frame” of the image so that it coincides with that star and so the others are “blurry”.
If your explanation is correct, then simple image overlays would be the explanation. However, stars are so far away that I do not see how any position change by the spacecraft could change their positions even slightly with respect the the star behind the target.
The explanation of tiny changes in the luminosity of each star makes more sense to me.
There is a fascinating article in Quanta magazine about extracting very faint light changes in images due to objects occluding light paths or acting as light sources. The algorithms to extract those images have many uses, including astronomy.
“I do not see how any position change by the spacecraft could change their positions even slightly with respect the the star behind the target”
I would state this differently. The proper motion of a typical star is often enough that it will move sideways (from our perspective) over one stellar diameter per day. The resolution of our optics is insufficient to resolve the motion because the image size is far smaller than a pixel. A far longer interval is required before the images don’t coincide.
If I understand you correctly, you are saying that any stellar drift is too small to resolve, so those artifacts are due to changes in luminosity, not position. Is this correct?
It should be computationally possible to eliminate those luminosity artifacts exposing the positions of truly moving, but very faint, solar system objects.
Sorry to take so long replying. On rereading the thread I see that I misunderstood the question, so my comment is poorly aimed. Let me try again. I speak of this as someone who has done a bunch of signal processing work though not on images.
First, there’s too much room for guessing since we don’t know how the image processing was done. Antonio may be correct in this guess on his point, but it’s difficult to say. The same applies to what I say.
The subtraction of two signals of similar magnitude can be extraordinarily challenging. The slightest mismatch, whether due to inherent instrument inaccuracy, raw data differences, noise or unexpected signal, and registration error, has an outsize impact on the calculation. The best result requires superb instrumentation and, very importantly, deep knowledge of the signal parameters.
Another source of error is that when the images are registered the pixel boundaries are unlikely to be perfectly aligned. That is, the pointing of the camera differing a tiny amount (of course, that why registration is necessary). The bleeding of the light into adjacent pixels will differ among images and can be very difficult to correct for. This would require phase information, not just magnitude, which the pixels cannot record: they’re just bit buckets.
So, artefacts are to be expected. Just getting the combined, subtracted image to this level of quality is impressive. But, per my earlier comment, it is unlikely the artefacts are due to proper motion of the stars. It may have an effect for some of the stars since (on rereading the notes) the images were taken one year apart.
That Edgar Allen Poe would have been so prescient as to predict in 1844 the existence of this KBO, by exact name for that matter, years before even Neptune was discovered, utterly boggles the mind. What other predictions of future unknown distant objects lie in dusty volumes of ancient rhyme?
Poe also “predicted” the Big Bang in 1849….
http://nautil.us/blog/edgar-allan-poe-part_time-cosmologistbig_bang-philosopher
“This first detection is important because the observations New Horizons makes of Ultima over the next four months will help the mission team refine the spacecraft’s course toward a closest approach to Ultima, at 12:33 a.m. EST on Jan. 1, 2019…”
“… will help the mission team refine the spacecraft’s course toward a closest approach … ”
ques.: HOW will this help the mission team refine the spacecraft’s course ???
http://www.hayabusa2.jaxa.jp/en/topics/20180828e/index.html
Have a look here for a very useful explainer.
(It boils down to comparing background stars and refining your model of the orbit to pin down exactly where the target is located).
As an amateur photographer, I am always amazed by the ability to aim at such a distant object and to obtain photos that are in focus. Could someone explain how targeting would be done in the Breakthrough Starshot program to achieve usable data from a much more distant object?
That one is the subject of intense speculation and the beginnings of a long line of research among the Breakthrough Starshot team. So at the moment there is no answer to give. Huge challenge.
“That one is the subject of intense speculation and the beginnings of a long line of research among the Breakthrough Starshot team. So at the moment there is no answer to give. Huge challenge.”
I don’t understand that ANSWER …
Breakthrough is at the beginning of the five-year period of its concept study. Tom is asking how a Starshot probe would target a planet around Proxima or other star to return good images, and so far we’re too early in the process to know the answer. I’m sure many ideas will be discussed and rejected along the way, but no single solution has been arrived at yet.
New Horizons Science Chat discusses Ultima Thule flyby
by Laurel Kornfeld
September 21st, 2018
NASA’s New Horizons spacecraft is now closer to its second target, Ultima Thule, than the Earth is to the Sun, and with the flyby just over 100 days away, mission scientists are actively preparing for the encounter.
In a Science Chat livestreamed on Sept. 19, 2018, NASA chief scientist Jim Green, mission Principal Investigator Alan Stern of the Southwest Research Institute (SwRI) in Boulder, Colorado, and Mission Operations Manager Alice Bowman of the Johns Hopkins University Applied Physics Laboratory (JHUAPL) discussed the upcoming New Year’s Day flyby of the small Kuiper Belt Object four billion miles from Earth.
Traveling at a speed of 31,000 mph (50,000 kph), the spacecraft will make its closest approach to Ultima Thule at 12:33 a.m. EST (05:33 GMT) Jan. 1, 2019, when it will pass within just 2,175 miles (3,500 kilometers) of the KBO, closer than it flew by Pluto in July 2015.
Ultima Thule and KBOs like it are made up of pristine materials from the earliest days of the solar system that became the building blocks of the planets, moons, and other small bodies orbiting the Sun.
“We’re going four billion years into the past,” Stern said. “Nothing that we’ve ever explored in the entire history of space exploration has been kept in this kind of deep freeze the way Ultima has.”
http://www.spaceflightinsider.com/missions/solar-system/new-horizons-science-chat-discusses-ultima-thule-flyby/