Dawn approaches Ceres, From opnav images to first orbit Options

[Bill Harris]

Quite a few hexagonal basins and craters out there...

But quite common. I recall hexagonal crater and lineations first being discussed in Ralph B. Baldwin's The Measure of the Moon and alluded to in his earlier The Face of the Moon.

And, as Algorimancer notes, there are many other studies on hexagonal structures.

--Bill

[Gerald]

Merged map based on the 27 RC2 frames of pia18920.gif:


Edit: Animation of maps by frame:

(size reduced)

Assumed globes, animated:

(half size)

[fredk]

It seems with every release we find something odd about the resolution of the images. I've compared the main bright spot image (PIA19185) from the previous release with frame one of the animation in the latest release. They're pretty clearly the same image, as you can see by comparing features near the terminator (that's why I've gamma tweaked these images).

However, the PIA18920 animation frame 1 is stretched vertically relative to PIA19185! For this gif, I've left PIA19185 alone (apart from gamma), but for PIA18920 frame 1 I've rotated 10 degrees, followed by a reduction in size to 90% (horizontally and vertically):

Previously we thought PIA19185 was supersampled 2x, so my suspicion is that the new animation frames have been altered. You can see a circular masked region on the new animation frames, with black outside. My guess is there may be some connection: someone tried to mask (I can't imagine why) with a circle and noticed Ceres isn't a sphere, so stretched it!

Anyway, this clearly means that anyone trying to create map projections from the images (Gerald, Phil) has to be careful.

[Gerald]

The mask is probably used to mask noise/dark frame outside the shape of Ceres.
Respective shape: I needed to adjust the diameter ratios to 920/975.
Images may be pincushion distorted a bit, but without knowing the position and rotation of the images, that's hard to infere.

[Phil Stooke]

The two opnav4 images show the north pole a bit better than before. Here I have added them to the northern hemisphere map.

Phil

[elakdawalla]

Previously we thought PIA19185 was supersampled 2x, so my suspicion is that the new animation frames have been altered. You can see a circular masked region on the new animation frames, with black outside. My guess is there may be some connection: someone tried to mask (I can't imagine why) with a circle and noticed Ceres isn't a sphere, so stretched it!
Anyway, this clearly means that anyone trying to create map projections from the images (Gerald, Phil) has to be careful.

Oh my gosh that explains the weird number of 230% of the original resolution for the vertical dimension. That is awful.

The question is, was the stretching done before or after the rotation?

Edit: after messing with it, it's clear the stretching was done after the rotation.

[elakdawalla]

I tried to match PIA18920 animation frame 1 to PIA19185 with stretching followed by rotation, but that isn't enough to make the two frames match.

[marsbug]

Nice bullseye pattern! If that's an old impact basin, it's diameter is either 380km or 630km, depending on which of two ridges is part of the rim.

The circumstantial evidence for ice volcanoes is growing. There's a diagonal fault line below centre, adjacent to which is an isolated mountain about 20km in diameter and 5km high, that just happens to be one of the bright spots. Then there's the bright spot on the prime meridian that has rays like an impact crater but the broad base of a caldera. It's near the terminator in the last image of the rotation sequence. No wonder the Dawn team is puzzled!

[attachment=35223:CeresBullseye.jpg]

Would carbon doxide outgassing from subsurface resevoirs, driven by solar heat, be possible in the Cerean temperature regime. We know (or strongly suspect) that this is the force behind the 'mars siders'...

[fredk]

I tried to match PIA18920 animation frame 1 to PIA19185 with stretching followed by rotation, but that isn't enough to make the two frames match.

Agreed, Emily, I also tried that. This makes it look like more was done than just rotation and vertical stretch - some weird geometrical distortion was also done.

I wonder if it might be that the two frames are not actually the same, but separated by very close to one Ceres rotation. That might explain some apparent geometrical distortion, but still would not explain the vertical stretch.

The mask is probably used to mask noise/dark frame outside the shape of Ceres.
Respective shape: I needed to adjust the diameter ratios to 920/975.
Images may be pincushion distorted a bit, but without knowing the position and rotation of the images, that's hard to infere.

But the noise is very low, as you can see by comparing masked and unmasked black. And a levels adjustment could have dealt with the small noise level more simply.

As for pincushion distortion, we don't have the full frames, but if, as I'd expect, Ceres is roughly near the centre of the frames, then I'd expect negligible distortion for a 222 pixel diameter region.

[Phil Stooke]

Possibly a reprojection from the perspective view of the camera to orthographic for mapping purposes?

Phil

[Herobrine]

A better combination of maps than I posted yesterday, now without 'ghost' duplicate features. Again, with Hubble color (also adjusted to align better).

This does not include any of the nice polar regions posted today.

[Gladstoner]

Are there any released images of the longitudes east of the bright spots? The global cylindrical map (pia18923) does show the area (in RC2 imagery?), but it has been shrunken down to 739x369 pixels for some reason.

Edit: Apparently not.

[Gerald]

Agreed, Emily, I also tried that. This makes it look like more was done than just rotation and vertical stretch - some weird geometrical distortion was also done.

I wonder if it might be that the two frames are not actually the same, but separated by very close to one Ceres rotation. That might explain some apparent geometrical distortion, but still would not explain the vertical stretch.

After 9 hours the distance of Dawn to Ceres will have changed, such that perspective distortion occurs. I've been considering distance since a while for the mappings, but not yet within one sequence.

[Gerald]

Are there any released images of the longitudes east of the bright spots?

I'm only aware of OpNav3.

[Gerald]

As for pincushion distortion, we don't have the full frames, but if, as I'd expect, Ceres is roughly near the centre of the frames, then I'd expect negligible distortion for a 222 pixel diameter region.

Whoknows; they are trying to minimize fuel consumption.