Dawn approaches Ceres, From opnav images to first orbit Options

[Habukaz]

Trying to bring out some detail in the images/animations posted thus far by increasing the contrast in some still frames:





First, some complex-looking sinuous details in the darker terrain surrounding the "bright" spot. It appears over several frames, so it is not noise (if it's real, I guess intersecting impact craters could explain it, rather than something more exotic; though this explanation looks less likely in the second frame).

Note how the bright spot is almost completely to the edge of the darker terrain in the above pictures, which is pretty strange. Maybe it is unrelated to the darker terrain? Though that would be strange, too; given that it's the only place that appears to have a terrain like that.



Second, what looks like a massive basin with a bright spot within.

Know I am waving my hands to the choir here but.... how Cool is this?
Here we are debating features on a world all of humanity is seeing for the first time. In all of human history
there is only One first time. And we are active participants. Thanks to this forum.

And come spring, we'll be doing it all over again with an even bigger planet.

I'll be honing my guess accurately what that heck we are looking at in these low-res images for the Pluto and Charon encounter on Ceres (even if they are very different worlds...).

The sharpness or blur of the frames is mostly due to artifacts introduced in the compression of the animated gif. The raw images would be much better quality, likely closer to that final frame. The public has not yet had access to those though.

Still don't quite see why only that one frame should be like that, though. Trivial detail, of course, but it intrigued me.

[Mongo]

i think you probably read this "Ceres is very dark and reflects only nine percent of the incident sunlight. The bright spots are brighter by approximately 50 percent – and thus as dark as an asphalt surface."

This would make the bright spots reflect about 13-14 percent of incident sunlight, which seems quite reflective for asphalt. A neutral grey card for photography reflects 18 percent of incident light, not much more than what the white spots reflect.

Fresh asphalt actually reflects about 4 percent of incident light, while old faded asphalt (which can be quite light in tone) might reflect up to 12 percent. So the bright spots would be between faded asphalt and a neutral grey card in reflectance.

[TheAnt]

A neutral grey card for photography reflects 18 percent of incident light, not much more than what the white spots reflect.

Fresh asphalt actually reflects about 4 percent of incident light.

Correct Mongo, I said grey in post 254.
I have no idea where this asphalt colour idea came from since it is one extremely dark material, comparable to charcoal.

[TritonAntares]

OPNAV3 hasn't been acquired yet. Imaging won't until later this evening, MST. From looking at Celestia, the observations times are (keep in mind that these are just the dwell times, the actual time frame when images are actually being taken is likely shorter and in the middle of these times frames):

Footprint 1: 2015-035 02:13-07:21 UTC
Footprint 2 (centered on Ceres): 2015-035 07:30-10:47 UTC
Footprint 3: 2015-035 10:56-14:35 UTC

The HGA is pointed toward Earth between 2015-0a35 15:53-2015-036 07:44 UTC (again, it might not be downlinking that entire time)

So maybe some images will be released Thursday, at the earliest?

Hi Volcanopele,
is there already any information about imaging session RC1 scheduled on Feb.12 available?

[Phil Stooke]

Well, RC is 'rotation characterization' - we get a full rotation this time.

Phil

[Gerald]

This is a refined - still intermediate - version using isometric projection/reprojection of the OpNav3 sequence (meaning proper rotation of the ellipsoid), and a gamma brightness stretch, based on pia19179-16.gif:
Longitude/latitude projection isn't yet quite accurate, since the distance to Ceres is assumed to be infinite. This leads to a motion blur in the merged image.
Assumed globe, animated lon/lat projections, merged lon/lat map, animated reprojection without, and with globe grid:

Next refinement step will consider the camera fov, hence finite distance, intended to remove most of the motion blur effect in the merged map.

Some other technical parameters for the projections used here:
Width of raw images: 800 pixels
Height of raw images: 360 pixels

Rough estimate of left border: pixel pos 347.0 (automated per image adjustment by object tracker)
Rough estimate of top border: pixel pos 135.0 (automated per image adjustment by object tracker)
Assumed diameter equator: 108.0 pixels
Assumed diameter pole to pole: 98.0 pixels

Tilt of Ceres rotation axis around x-axis: -23.0 degrees
Tilt of Ceres rotation axis around z-axis: -9.0 degrees

Longitude rotation per image: 6.0 degrees

Weight used for merge: 1 - r, with r the relative distance from the center of the underlying sphere
Brightness stretch: quadratic (biquadrated raw values)

[Bjorn Jonsson]

Here is an image from a stack of the last 6 frames, sharpened with an unsharp mask and corrected for Ceres' rotation. I used Emily's GIF animation from here since the images have been aligned more carefully than in the original GIF.



This doesn't add much details but possibly some. One problem is that the interval between images is approximately three times longer in the OPNAV3 animation than in OPNAV1 and OPNAV2. This makes it impossible to stack as many images with good results as in the earlier OPNAVs.

It now looks even more like there is a fairly big crater with a cental peak not far from the terminator and slightly left of center. But it now looks as if that crater might be inside an even bigger crater, apparently slighly SW of the big crater's center. There is probably a smaller crater WNW of the central peak. It is still far from clear to me whether Ceres has something resembling Ithaca Chasma. I had expected this to have become clear at this resolution but it isn't.

I also took a quick look at a 'nearby' icy satellite, Callisto, at comparable resolution (~15 km/pixel). The contrast between bright craters and other parts of the surface is much greater on Callisto than Ceres. Ceres' contrast is more similar to Saturn's satellites, at least at this resolution.

[Gerald]

I'm not yet quite satisfied with retrieving camera parameters (fov, positions, etc.). But some features seem to get more clear already.
This image is derived from the frames 8 to 20 of the file pia19179-16.gif (preliminarily map-projected, averaged, and stretched):

My impression is, that the radial features around the bright spot (in the left upper part of the image) resemble a system of (dendritic) valleys.
The image isn't sharpened, therefore the brightish seams around the radial structures don't look typical for impact ejecta; melt-up might be an option.
Some other bright features (e.g. a little right of the center) may be (at) slopes of valleys/craters.

[dvandorn]

I've been looking for any sign of rays from the White Spot, and on the latest opnavs, I do believe I see some.

There is a fan of higher-albedo terrain splayed out over an arc of a good 30 to 40 degrees, from almost directly west to west-northwest from the spot. It's subtle in the latest opnavs, but I am convinced it's there.

Looking in the other direction, to the east and especially northeast, you see the dark terrain, but again in the latest imagery I begin to see something of a wedge-shaped structure to it as it comes out of the spot.

So, while I'm not saying anything definitively, I would not at all be surprised to see, when we get closer, a big ol' crater, very much brighter than its surroundings, with light-toned ejecta to the northwest and dark-toned ejecta to the east and northeast. If so, that thing is going to be very, very impressive to look at.

-the other Doug

[Bjorn Jonsson]

Maybe - but e.g. Callisto has lots of craters that are much brighter than their surroundings even though they have no rays (or only subtle ones). Here is a view of Callisto at comparable resolution to the latest OPNAVs, i.e. ~15 km/pixel.



A higher resolution example of a bright crater on Callisto that doesn't have any rays can be seen here: http://photojournal.jpl.nasa.gov/catalog/PIA01128

[nprev]

Ah. Thanks, Bjorn; I knew that Ceres was looking sort of familiar in this regard.

Can only take the Callisto analogy so far, though. Ceres is already showing large craters and significant surface deformity & compositionally it's probably not very similar in terms of rock/ice ratio. Based on that I'd guess that we probably will see at least faint rays around the bright spots at closer resolutions.

[4throck]

A quick overlay of Hubble color over one of the latest images.
There seems to be some correlation.




Here the same exercise combining Phil (post 130) and Gerald's (post 276) maps:

[Gerald]

In the meanwhile I've pinned down a set of parameters which returns for frames 6 to 20 of the OpNav3 sequence pia19179-16.gif results sufficiently accurate to try a next analysis step.
I'm using this globe for the projections:

(Using actual distance (146000.0) and size data (975/909 km diameter) for Ceres, 5.8° rotation steps between images, Ceres axis tilt around x-axis -15°, around z-axis -10°, apparent equator diameter 108 pixels.)
First, if you look at the full sequence (square encoded), you may notice, that the right-most 5 frames rotate:

Therefore I'm only using frames 6 to 20 to get a weighted average:

Since the human eye can also merge a rapid sequence of images, here a fast animated gif of the subsequence:

You may note, that the brightness of nearby pixels can vary differently over the frames. This can indicate different slopes, more formally different surface normal vectors.
Different brightness can also indicate different albedo.
Together we get one of the simplest reflectance models, the Lambertian or diffuse reflectance model.
In the case of just one color (grey) and one light source (the Sun), the diffuse reflection is described by just three variable real-valued parameters: albedo (color), and two angles to describe the surface normal relative to the Sun; the Sun intensity can be considered as constant.
Actually the two angles can be reduced to one angle relative to the Sun, since the model is symmetric to the pointing towards the Sun. The intensity (brightness) of the pixel is proportional to the cosine of the angle.

If we now follow one pixel over several frames, the angle of the surface normal relative to the Sun changes, resulting in varying grey values as a function of the frame number.
The model refers to linearized intensities. So instead of the above square-encoded frame sequence, the linearized version is needed:

Raw images are usually square-root encoded; therefore I've squared the pia19179-16.gif images to get the linearized version. If they are gamma-corrected with gamma = 2.2 for standard displays, the result is rather similar.
Merge of linearized frames 6 to 20 by weighted averaging:

The first of the following two diagrams shows the grey scale of one pixel of the longitude/latitude map as a function of the frame number.
The second diagram shows the grey scale function of the 720 pixels of one longitude over all latitudes (in 0.25 degrees steps), meaning one column of pixels:

(For the diagrams here, the origin of the lon/lat maps is the lower left corner.)
Text version for this considered column:
 pia19179_16_xx_rectifieda_098_grey_col693.txt ( 78.75K ) Number of downloads: 721


Many of the curves in the diagram are rather noisy, others appear reasonably smooth. The jiggery ones are probably mostly due to Moirée-like artifacts near the border of the projected area.
The next days I'll try to match the reflectance curves of each pixel position via RMS minimization to the simulated reflectance curve of a slope on the surface to retrieve albedo/slope data for each pixel.
If the anticipated result won't be too noisy, it may be usable as an albedo map, and as a basis to infere a topographic map.
If the data turn out to be of good quality, residuals may provide shinyness information, according to the more general Phong reflectance model, usable as an indicator for surface roughness.

[Ian R]

This is fantastic, Gerald, I'm really learning a lot here. I presume you will be applying these techniques to Pluto, once New Horizons begins to resolve surface features?

[Gerald]

You're quite right. I'm working on improving my capabilities also for New Horizons/Pluto and Junocam next year.
Some others may add their capababilities, too, and we'll be well-prepared for those missions.