Obtaining the 23k Viking merged color/MDIM 1.0 map

[sombrehombre]

Has anyone had any luck finding a contact (David Seal at JPL, or anyone
at USGS Flagstaff) that can provide access to this source material.

I've sent a volley of emails to any reasonable contact I can find regarding
these with no luck so far.

Thanks
Darryl Roy
sombrehombre@hotmail.com

[selden]

Darryl,

Have you tried asking the USGS planetary map server to create a map that size?

http://pdsmaps.wr.usgs.gov/PDS/public/explorer/html/marsadvc.htm

In principle you can use the form to tell it to create an image with 64 pixels/degree, latitude box= 180, longitude box = 360.

It'll complain that the file is too big to provide in a Web page, but that you can "order" it. That just means the file will be written to an ftp directory for you to download. It'll send you e'mail when the file is ready. It does not mean that you have to contact someone to ask for it.

I've used this method to create smaller maps with no problems, at least.

[sombrehombre]

I've placed orders for full resolution maps, for hemispheres, and for semi-hemispheres, in .tiff, .raw, and even .jpg format, with no luck. As far as
I can tell, the automated system is rejecting my inquiries as being too
large, and I'm getting filed in the junk mail folders of all the email contacts I've sent inquiries to.

David Seal at JPL once had this map up on the map page, but its now
listed only by request. He, too, hasn't replied to exceedingly polite
inquiries at his samhadi domain email address.

[jestr]

I think you can find it here
ftp://webgis.wr.usgs.gov/mars/usgs/
I just downloaded a 23k JPG (not very well put together) from this site and it looks to have other useful Mars maps also.Jestr

[sombrehombre]

You rock - I had tried to obtain access to this site before, but it was down.

The Viking map is a merge of a medium resolution color map, made from
the red and purple Viking orbiter filter mosaics, with a synthetic green
channel, and a higher resolution clear/bw channel. As far as I know, its
the only high-res global mosaic that has at least two real filter channels in
register (MOC mosaics are red only, with blue channel used for cloud
observation - Odyssey/THEMIS images are plagued with lightleaks, so
only two color images have ever been published).

Rolling your own green channel from the red and blue channels does
wonders to eliminate the wholly unrealistic yellow (burnt) highlights
and purple (dodged) dark sands. A contrast adjustment fixes the
contrast added by USGS to highlight albedo & topographic features.
From there, its simple work to adjust the color pallette to match true
colors seen in Pathfinder, Spirit, Opportunity, or Mars Express images.

Whenever the Mars Express global mosaic becomes available, it will put
everything else to shame - it has the best color rendition of any orbital
platform yet. Go Europe! But since thats at least a year or two off, I
thought I might spend an evening making my own Mars texture.

I doubt any useful work for Celestia will eventuate from this - there
are already superior 32k flat & normal VT maps available. I mostly
wanted a copy because I hate the colors of the Space-graphics red mars,
and wanted to make the definitive 16k preshaded Mars for Orbiter.

[jestr]

Sombrehombre-how do you 'roll' your own green from the red and blue channels.I think I can do the rest of it.I'm using Photoshop,Jestr

[sombrehombre]

A lot of planetary imagery is taken with filters that don't correspond to
human eyesight - Voyager color images are frequently 2 color images
in orange and violet, with synthetic green. The Clementine based maps
that you've seen before are green because they mapped blue to blue,
near-IR to green, and far-IR to red. In the case of Viking, the original
colors were violet and red, and for the Viking based color maps, a synthetic green based on a weighted 'average' of the two was used. This leads to really weird purple 'dark' sands (where the green channel was 'dodged') and bright yellow (where it was burnt). Additionally, they added tons of contrast to highlight the difference between light dust and dark dust (which comprise the most prominent albedo features).

Its not exactly true color.

Your own weighting is equally valid (ie, its down to aesthetics).

To make your own green, based on a plain 'weighted' average, just decompose into R, G, B greyscale channels, discard the G, duplicate R or B, and paste the other on top (of the duplicate) with a varying transparency. The G will then simply be a value uniformly intermediate between the R and B. Recompose the R, B, and new G, and then you can play with fine tuning with color curves and the like.

This can be done in either NetPBM (arithmetically) or GIMP (graphically). I've never found a need for Photoshop.

The derived color formula could be made arbitrarily complex (there's 256^3 possible mappings with 16->24 bit color). This one, which is more or less like a photographic duotone, appears to work well since aside from the Earth and Io, spectrally sharp green/yellow compounds are rare on planetary surfaces. For my moon, an unweighted average with no post-processing of the Clementine color channels gave an exact match to the hue (though not saturation) seen in reddish and bluish maria, and tannish highlands with 3 color CCDs on Earth telescopes.

[jestr]

Thanks mate,Jestr

[sombrehombre]

With that Viking color map, I'm having a hard time improving
much over an unweighted (set transparency/blend to 50%) average
of red and blue channels, followed by a linear color remapping in
each of the three channels, with the color triplet
(0,0,0)->(68, 58, 68) and (255,255,255)->(255,255,255). It lacks
a little in shadow contrast (AND ISN'T ALIGNED WITH BUMPMAPS!), but appears to have the color and albedo range well matched to the 1995 3-color Hubble images.