[X] My Assistant

[Bjorn Jonsson]

After extensive software improvements to speed things up I just finished doing an experimental Iapetus animation based on a new and bigger version of the DEM described in this thread. The new version looks like this:



The animation can be downloaded from here (almost 5 MB). It requires the DivX codec (see http://www.divx.com )

A sample frame from the animation:


The animation starts and ends 3000 km from Iapetus' center and gets to within about 300 km from the surface. Vertical relief is exaggerated by a factor of 3. I may do a new version later with the horizon horizontal at close range and not vertical as in this animation. Flying over a bigger part of the surface at close range is another idea.

The texture map was generated as a byproduct when doing the DEM so it fits the DEM perfectly. It was processed to make dark terrain lighter, otherwise the animation appears too dark.

Once I have a DEM covering a bigger part of Iapetus I will update the animation.

[djellison]

****ing hell that's good. This community is lucky to have code monkeys like you

[nprev]

Uh....WOW?!?!?!?!!!

I was gonna try to inject a creative curse, but words honestly fail me. Thanks so much for this, Bjorn!

[tedstryk]

Wow, that is amazing! It looks very real - not the cartoonish feel that such things have when based on poor quality DEMs.

[mchan]

Ditto Wow!

[djellison]

not the cartoonish feel that such things have when based on poor quality DEMs.

Ahh - so you HAVE seen the HRSC stuff then.


Doug

[ugordan]

This has got to be one of the coolest animations that came out of Cassini, ever. It's the next best thing to actually having a shape model of Iapetus on your desk!

Outstanding work, Bjorn!

Me wants a global DEM at that resolution... *droool*

[ngunn]

I'd love to see this (even with 3x vertical exaggeration) but unfortunately it won't play on our system.

As a matter of curiosity how do you define vertical exaggeration on a non-spherical body?

[Ian R]

Bjorn, you're going to have to put this animation on YouTube - it's simply awesome!

[ugordan]

Nigel: You can calculate normals to each point of the surface and exaggerate along that direction, which works fine even for triaxial ellipsoid approximations.

[ngunn]

Nigel: You can calculate normals to each point of the surface

Yes, but those normals won't be truly vertical in a gravitational sense. I'm sure the difference is reasonably small for Iapetus, but it might be considerable when we get to somewhere like Vesta. If you wanted to compare local slopes with the angle of repose (for example) you'd have to be careful with this. Of course you could avoid the issue completely by doing no exaggeration.

Anyhow, I'd better leave this thread to those of you fortunate enough to be able to view the animation.

[ugordan]

If it's gravitational-sense-vertical you're worried about, you can compute the center point of the body (center of the ellipsoid) and use the surface radius vector instead of surface normal vector for exaggeration. That should be damn close to gravitational vertical.
I wonder which method Bjorn actually uses. From a rendering point the surface normal method appears to be more straightforward.

[Bjorn Jonsson]

The vertical exaggeration is very simple. The DEM is normally stored as lat,lon,distance from Iapetus' center. It's also possible to output it as lat,lon,deviation from a triaxial ellipsoid. I simply multiply this deviation by a factor of 3.

Me wants a global DEM at that resolution... *droool*

This is exactly what I want - for each and every icy Saturnian satellite. I'm hoping to get at least 50% coverage for Iapetus once the September 2007 flyby images hit the PDS. It must be possible to get much more than 50% coverage for Tethys, Dione and Rhea. Mimas is rather poorly imaged relative to the other satellites (especially near and east of Herschel) but that should improve greatly with the 9500 km flyby in 2010. Enceladus is more difficult because of its smoothness.

A DEM for every satellite is a huge project. I've been wondering about possible 'collaborative DEMs' recently. In particular, measuring control points is a lot of work. The process could probably be partially automated but doing so is a lot of work too.