Thanks Emily!

Here some further brightness stretch and adjustment, cropped:
Click to view attachment

I think, that's now close to the limits of what can be done with "basic" image processing techniques.

False color animated gif of the dancing moons:

(background blue, very bright variable parts red, subtle objects greenish)

Brightest stars (not too close to Pluto/Charon), and the presumed Pluto/Charon barycenter with crosshairs, cropped:


Background masked, to show only the above marked stars:
Click to view attachment

List of star pixel positions in the above graphics, and integrated brightnesses:
Click to view attachment

The brightest object should be close to the presumed Pluto/Charon barycenter.
This position is useful to narrow down the parameter space where to look for other moons, and to pin down the orbital parameters of Nix and Hydra.

Starting with uploads of two processed images to astromomy.net (and this image) following a hint of Emily's very useful TPS blog post, I've obtained the coordinates of the star background.


Then I've the assembled a blink gif between a map obtained from Google Sky as a reference, and a roughly brightness adjusted version of the filtered and merged OpNav Campaign 3, LORRI 4x4, images (until 2015-05-01):


Besides the bright Pluto/Charon barycenter, I found two subtle, but significant differences:
1 - could be a glitch in the reference map, or a variable object,
2 - unclear, could be a variable star, or a TNO / moon; I can rule out a processing artifact almost with certainty, as well as an incompletely filtered CR hit, since the feature is present in several LORRI images.

Here's my take on moon IDs. Sigma clipped & aligned the last four quints in Maxim DL. No sign of Styx or Kerberos yet but it shouldn't be much longer. Click for animated GIF:

Click to view attachment

Great!
I get more and more the impression, that Nix and Hydra are rather non-spherical, and rotating.

Nonspherical is obviously highly probable for both moons, and of course they're rotating. Be interesting if they're not rotationally locked with respect to Pluto, though, or if there's some sort of odd harmonic relationship as we've seen elsewhere.

speaking of which, are there any ideas of the rotation status of the minor moons? I guess a classical spin-orbit resonance would be difficult to achieve in such a binary system, as it would be like orbiting around a highly non-spherical primary. would a spin resonant with Pluto's spin and Charon's orbit be more likely?

I think a standard spin orbit resonance is still most likely, with perhaps larger physical librations than we are accustomed to. The Saturn system is very complicated, but I think only Hyperion is in an unusual rotation state.

Phil

Well, Mark Showalter at least thinks they could be chaotic rotators, but not everyone is convinced.

As we understand it, Pluto's surface is a mixture between bedrock and N2, CH4, CO and C2H6 ices cover, plus the atmospheric gasses freezing on the surface (N2, CH4, CO, Ar, higher hydrocarbons (acetylene, diacetylene) and nitriles (hydrogen cyanide, dicyanoacetylene) ), so the situation is much more complex than just simple nitrogen (N2).

By the way, here is some info on Pluto's and Charon's Sun occultations, in case someone hasn't posted them already.

"For both ingress and egress, far-ultraviolet solar occultations will be observed with the Alice instrument (wavelength bandpass: 52-187 nm) and uplink radio occultations will be observed with the REX instrument (X band wavelength: 4.2 cm), for both Pluto and Charon. The solar and radio occultations mostly overlap in time (since the Earth and Sun are very close together (0.24°) as seen from Pluto); the Pluto occultations are about an hour after closest approach (e.g., we will observe the solar ingress at Pluto from 12:16-12:49 UT on July 14th, and expect the Sun to set at about 12:47 UT; closest approach to Pluto is expected at 11:50 UT). The Charon occultations occur about an hour after the Pluto occultations. Besides all these occultations (we also observe the occultation by Pluto of a bright UV star with Alice, starting about 4 hours after closest approach), we use the Alice instrument to look for “airglow” (a faint glow at certain UV wavelengths, mostly due to processes related to UV sunlight; e.g., a UV photon from the Sun ionizes an N2 molecule, the emitted electron may be energetic enough so that when it hits another N2 molecule it breaks it apart and excites one of the N fragments to emit a photon at 134 nm). The particle instruments will indirectly study atmospheric escape by measuring ions formed when solar wind protons collide with escaping N2 molecules and steal an electron – the N2+ ions are swept up into the solar wind and are called “pickup ions”. We’ll also use the LORRI and MVIC imagers to look for clouds and hazes, or even plumes (as were seen at Triton, a near twin of Pluto)."

From an email conversation I had with Randy Gladstone, atmospheric lead scientist on New Horizons.

I was talking about the polar ice cap specifically and for what it's worth, I was quoting the article. Solid Nitrogen and Carbon Monoxide are perhaps more likely to be transient. Methane is more likely to be constantly frozen with perhaps a smaller proportion in the lower atmosphere. The remainder including Argon, higher hydrocarbons and of course thiolins would be almost permanently frozen. Hydrogen cyanide would have negligible vapour pressure and may be likely to be found in association with thiolins in more ablated regions, although over time it tends to hydrogenize even in the solid phase. I would add methanimine (CH2NH) to your prospective list.

Either way, a lot of this is speculation, but we only have to wait a couple of months to find out more.

Starting from manually determined positions of Nix and Hydra

in 17 (until 2015-05-06) of the OpNav Campaign 3, LORRI 4x4, cleaned, stacked and star-background-subtracted image quintuples I've tried to simulate the positions by manually adjusting parameters of assumed circular orbits:

This simulation can be extrapolated:

After narrowing down the orbital plane I've transformed the processed LORRI images, such that objects orbiting in a circular way around the Pluto/Charon barycenter are mapped to almost constant position; phase angle as x, radius as y-axis:

Here an example of the last eight of this image sequence stacked to reduce noise:


Versions on less

and more processed images:


For small radii the method is very sensitive to the barycenter.

Nix and Hydra are apparent, Styx and Cerberus not yet evident (to me), despite this excessive processing.

I agree, it's definitely speculation, but it's not like we can do much more before July 14th Yes, the purpose of my comment was just to say that reducing it to one element may lead you to misunderstand the broader picture. Your comment is absolutely fine! Plus, like you say, with the few data we have it's impossible to say anything with certainty. Can't wait for July!!

Stellar image processing, Gerald. Really enjoying your work.

I just noticed that the image metadata files have the expected X,Y coordinates of all the objects in each raw image. For example:

http://pluto.jhuapl.edu/soc/Pluto-Encounte...0x633_eng_1.txt lists:

PLUTO;X=127;Y=126
CHARON;X=128;Y=137
NIX;X=119;Y=99
HYDRA;X=116;Y=156
KERBEROS;X=152;Y=122
STYX;X=107;Y=126

These seem to have the Y origin at the bottom of the image, and I needed to subtract a pointing offset from all of them equally to get things to line up. But this could help with spotting Styx and Kerberos, perhaps doing a manually shifted stack on the expected XY. Strip the URL back to /info/ for a directory of the files.

Finally, Kerberos and Styx!

http://www.nasa.gov/image-feature/nasa-s-n...est-known-moons

Planetary, actually.

All five this early, awesome.

That is all the known moons down. Now we keep watch for the unknowns (if they exist).

First congrats to the New Horizons team, and particularly to John Spencer and his image processing team, for accomplishing this milestone!
At least with the published jpegs until 1st of May that's exceedingly difficult.



The coordinates seem to be relative to the barycenter of the system.
I tried to adjust for this by adding the displacement needed to register stars:
Click to view attachment
But there is an increasing error build-up. I've been pondering for a while for the reason; my best explanation thus far is a changing parallax of the Pluto/Charon barycenter with respect to the starfield background during the OpNav Campaign 3.

Is the error build up due to the increasing apparent diameter of the system as NH gets closer between frames?

The decreasing distance (radial velocity) seems to be considered.
My best guess is, that the shift is caused by the proper motion / transverse velocity (applying the astronomical notions for velocities of stars by replacing the solar system barycenter by New Horizons).
For a more accurate modelling the transverse velocity components of Pluto/Charon and New Horizons need to be considered.
Previously I've assumed the transverse velocity as zero/negligible.

NH raw images (binned) with timestamp today have been released, which is interesting considering the fact that images at these distances have a significant probability of containing hitherto undiscovered/unpublished moons.

All the obvious geometrical effects are included in the predictions of the positions of the objects in the images- this is all done with SPICE which knows how to make those calculations. The offsets are mostly because the spacecraft orientation is only determined to within a certain precision, that is bigger than the resolution of the camera, so you can't predict the exact pixel objects will land on from the spacecraft pointing data alone. In fact the first thing we do when analyzing the images is to determine accurate pointing by comparing the star positions to a catalog. I'm not sure why the error would show systematic drifts with time, but it might be just coincidence.

John

It seemed to me to be a drift of the Pluto/Charon barycenter relative to the star background, somewhere near 3e-3 degrees over more than three weeks.
I could imagine a slightly curved trajectory of New Horizons due to the Sun's field of gravity, or an effect of Pluto's orbital motion, or just some glitch in my processing or perception.
But that's just preliminary ideas; I haven't done a fully accurate analysis (yet), and may post a revision later.
(I'm intending to take a look at the hazmat images, first.)

For an accurate analysis of the orbits of the moons, however, it will likely be necessary to determine or rule out the presumed drift.

To a first approximation, the spacecraft and the Pluto-Charon system are approaching each other along the unequal legs of a right triangle, so why shouldn't the barycenter move with respect to the background star field? The barycenter is not the bulls-eye of a stationary target.

I've been straining my brain over tanjent's idea. In rough numbers, Pluto is moving at about 5km/s towards the point where closest approach will occur. NH is moving towards the same point at about 14km/s. Imagining the third side of the triangle as the line joining NH to Pluto, the three sides will shrink in proportion to each other so that the direction of the line joining NH to Pluto stays approximately constant. I say approximately because the paths are curved, they aren't really aiming at exactly the same point etc. Imagining myself riding along with NH, I'm not surprised that Gerald finds a very small drift for the barycentre - the apparent motion of Pluto relative to the background stars as seen from NH would be much less than that seen from the Sun.

When I first looked at the new processed images from the NH spacecraft showing the motion of all four of the small moons, I couldn't help but notice that there was a diffuse blob that starts out beside Hydra and is apparent in the last 3 of the frames. Let me just say I understand this is probably just an artifact. But I haven't quite been able to convince myself of this so I started wondering what if it was real. So in the interest of mostly just pure speculation, what if it is an ejecta plume from an impact event on Hydra (I know what are the odds that we would just happen to observe this). I'm not familiar enough with the orbital mechanics of this system to try and say if the apparent motion of the object would even fit such a scenario. Thoughts?

http://www.nasa.gov/image-feature/nasa-s-n...est-known-moons

Yes it makes sense to have a small drift from the curvature for NH's and Pluto's paths. This seems consistent with being a few percent of the drift of Pluto in the sky as seen from the sun, given NH's speed relative to Pluto. Unless Gerald is referring to an error in the drift vs the drift itself. If Celestia or similar software has NH added to it, then this drift can be independently evaluated?

When you push the image processing as hard as we had to do to see Styx in this sequence, all kinds of artifacts start to emerge- that fuzzy blob is probably some sort of internal reflection. We could have manually cleaned up the release image to remove everything we weren't sure was real, but we decided a "warts and all" release would give a better idea of what the data are really like, and the challenges of working with the images (plus, we don't have that kind of time ). We'll have much better images soon, of course.

John

I believe each of the four smaller moons is now 1 pixel wide and was teased out of background noise. The blob you see is much less than the original 1 pixel, so it's just an imaging artifact.

Also don't forget that the spacecraft had a different attitude relative to the Plutonian system during some of these images so the direction of the artifacts may change with each image.

Andy

P.S. Sorry John for the duplicate. I started typing just as you hit send.

Here a stitch of seven 2015-05-11 "Search for sources of hazmat" 6-image stacks:
Click to view attachment

It has been rather easy to identify Nix and Hydra. But I wasn't able to find Styx and Kerberos. Here a crop of a heavily processed version, obtained by subtracting an appropriately brightness-stretched version of the OpNav Campaign 3 stitch, and in parallel subtracting a processed Google Sky image roughly simulating the LORRI images, then combining the two differences to filter out stars as far as possible:
Click to view attachment
I've tried several processing techniques, but none was successful in revealing the two smallest known moons of the Pluto/Charon system.
I guess, that the data loss due to jpeg compression and reduced bit depth has swallowed the tiny signal of the two moons.
Some discrepancies (pointed out in some preceeding post) to Google Sky have been reproducible.

Some images of OpNav Campaign 3 aren't yet included in my processing. Might be, that adding the latest images of the sequence, and considering the potential transverse velocity component will add up to a perceivable signal.
I hope, I'll be able to investigate this over the week-end.

"Search for soures of hazmat", stitching by median-filtering, blink gif

and differencing of cleaned and summed images:
Click to view attachment

Some more elaboration of the presumed parallax:
A two-image blink gif consisting of two cleaned, stacked, registered and cropped OpNav Campaign 3, LORRI 4X4, images, taken with an interval of about 32 days:
Click to view attachment
I'd say, it shows the parallax of the Pluto/Charon system evidently with respect to the background stars.
The increasing apparent angle between Pluto and Charon is evident, too.

21 sets of OpNav Campaign 3, LORRI 4x4 images, source images and stacked version, synopsis:


Edit: Album of registered versions of the stacked images.
Reference point and rotation angle for respective applied (north-south aligned) registration: Click to view attachment

A few days from now New Horizons will obtain unbinned images that will have almost two times better resolution than the images released in April. In these images Pluto will be a whopping 8 pixels or so across - I may never have been looking as much forward to seeing such tiny images as I'm doing now. These images might even start revealing something on Charon if it has large scale markings.

I decided to assemble a quick montage comparing the best released NH Pluto images so far to several outer planet satellites at low resolution. For the satellites I used Voyager images except at Saturn where I used Cassini images. I used two images of each body. In the best NH images so far Pluto is roughly 8 pixels across following stacking and extensive processing (in the original images it is 4-5 pixels). I decided to make the satellites bigger, about 12 pixels across, mainly because soon we'll be seeing better NH images than the ones released last month. In all of these images where possible the phase angle is comparable to the phase angle in the NH images. I made no attempts to correct the image orientation (i.e. make north up).

This is the result:

Click to view attachment

Pluto is known to be a high-contrast body. It is obvious from this (with lots of caveats since I don't know the details of the Pluto image processing) that the contrast is rather big but not nearly as big as on Iapetus. The best satellite analogs in terms of contrast seem to be Dione, Ganymede, Io and maybe Ariel. Triton exhibits far less contrast than Pluto. It's interesting to note that Ganymede has polar caps that are brighter than the equatorial regions.

It's going to be really exciting to see the true nature of Pluto's bright/dark markings revealed over the next several weeks.

And here is a version of the above image where I have attempeted to correct for the albedo of the bodies. This is only approximate and should be taken with a grain of salt.

Click to view attachment

In my latest "OpNav Campaign 3, LORRI 4X4" series I've chosen south up (Astrometry.net) (within 0.1 degrees accuracy, most likely according to the J2000.0 system) to stay compatible with google sky, and to simplify comparison with a generally accepted frame of reference.

Regarding FOV: The new-image.fits (link on Astrometry.net site) image header says 1.02179 arcsec/pix (using the ESA/ESO/NASA FITS Liberator 3), that's 1024 x 1.02179'' / 3600(''/°) = 0.29064° fov for 1024 pixels.
(I've submitted an average image, obtained from the 21 cleaned, 4-fold magnified, stacked and registered 4x4 binned LORRI images; the pixel scale is hence for the 1x1 binned images.)

New images from early May have been released!

http://pluto.jhuapl.edu/News-Center/News-A...p?page=20150527

Wowzers at the latest images! I made a quick animation with the images closer to native resolution.

Click to view attachment

Does anyone have the specs of the Pluto and Charon occultations that NH will use to study their atmospheres? Not really sure if this is the correct place to ask it, please excuse me and feel free to move them somewhere else if necessary.

Sure.

For both ingress and egress, far-ultraviolet solar occultations will be observed with the Alice instrument (wavelength bandpass: 52-187 nm) and uplink radio occultations will be observed with the REX instrument (X band wavelength: 4.2 cm), for both Pluto and Charon. The solar and radio occultations mostly overlap in time (since the Earth and Sun are very close together as seen from Pluto); the Pluto occultations are about an hour after closest approach (e.g., NH will observe the solar ingress at Pluto from 12:16-12:49 UT on July 14th, and the Sun is expected to set at about 12:47 UT; closest approach to Pluto is expected at 11:50 UT). The Charon occultations occur about an hour after the Pluto occultations.

Besides all these occultations (NH will also observe the occultation by Pluto of a bright UV star with Alice, starting about 4 hours after closest approach), NH will use the Alice instrument to look for “airglow” (a faint glow at certain UV wavelengths, mostly due to processes related to UV sunlight; e.g., a UV photon from the Sun ionizes an N2 molecule, the emitted electron may be energetic enough so that when it hits another N2 molecule it breaks it apart and excites one of the N fragments to emit a photon at 134 nm). The particle instruments will indirectly study atmospheric escape by measuring ions formed when solar wind protons collide with escaping N2 molecules and steal an electron – the N2+ ions are swept up into the solar wind and are called “pickup ions”. NH will also use the LORRI and MVIC imagers to look for clouds and hazes, or even plumes (as were seen at Triton, a near twin of Pluto).

Source: a conversation I had with Randy Gladstone, who leads the atmospheric studies for the New Horizons mission team.

Very nice Pluto images - hard to believe. In comparing the contrast to other satellites Earth's moon comes to mind.

Randy Gladstone, 'Gladstoner' - any relation? I always had the latter down as a geologist.

Here are the three images side by side. Looks mappable!

Phil

Click to view attachment

It's good to see the surface features of Pluto gradually taking shape.

Hang on. Is it me, or just pareidoglia ?

Incredible new images. I mentioned new images in my montage post above but the images I had in mind were images that will be obtained at the end of this month. I was expecting them to look similar to these new images so now I'm getting even more interested in seeing them.

Looking at the three images posted above by Phil it's starting to look as if there might be fairly strong longitudinal contrast variations. There seems to be considerably stronger contrast in the leftmost image but this may be a processing artifact. Anyway I'm now getting seriously interested in knowing which of the features (or which hemisphere) visible in the new images are going to be imaged at really high resolution near closest approach.

Had to do this...
Click to view attachment

Those albedo differences! Seriously exciting.

Pluto looks less spherical than I expected.