Printable Version of Topic

Posted by: AlexBlackwell Jun 14 2006, 02:03 AM

I wasn't sure where to put this but, assuming it hasn't been mentioned already, some of the image processing gurus here may be interested in a paper currently in press with Planetary and Space Science by Lohse et al. entitled "http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6V6T-4K5ST7F-1&_user=10&_handle=V-WA-A-W-AV-MsSAYWW-UUA-U-AACWYVAUBA-AACUVWAYBA-EBCBCEDYB-AV-U&_fmt=summary&_coverDate=06%2F13%2F2006&_rdoc=1&_orig=browse&_srch=%23toc%235823%239999%23999999999%2399999!&_cdi=5823&view=c&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=c9bdbb66d8233a957ca99749d04a36f0." I found what appears to be a http://www.isprs.org/istanbul2004/comm4/papers/458.pdf from the http://www.isprs.org/istanbul2004/, which was held in Istanbul, Turkey, in 2004.

Posted by: DonPMitchell Jun 14 2006, 03:51 AM

Tony Cook did work about six years ago on stereo topography from Clementine images. Odd that the new paper does not cite their work. http://www.cs.nott.ac.uk/~acc/dems.html. Note the maps of the north and south poles.

Posted by: DDAVIS Jun 14 2006, 01:19 PM

I keep waiting for the global Lunarf DEM derived from this work to be released.
Don

Posted by: DonPMitchell Jun 14 2006, 06:06 PM

I think one problem is the overlap for good stereo imaging is very spotty, only really good at the poles. I'm waiting for good Lunar altimetry too, for my solar-system simulator. I think it just won't exist until the LRO mission.

I haven't looked into what SMART-1 is doing. If its images were available, would people be able to do multi-image analysis?

Posted by: Phil Stooke Jun 14 2006, 06:26 PM

Tony has let me use a very low res version of a dataset consisting of his stereo and the Clementine altimetry. I think he and his coauthors are still working on papers from his DEM before he releases it, but I do think it will appear at some point.

Phil

Posted by: djellison Jun 14 2006, 06:30 PM

I've seen some photoinclinometry work from the new HiRISE stuff, and other stuff even on this forum - does any one know of any open source or freely available software that could have a hack at that?

Doug

Posted by: DonPMitchell Jun 14 2006, 08:05 PM

I'd be surprised is there is good canned software for this, especially tuned to the bidirectional reflectance of the Lunar surface. It's almost easier to just write something in C++ that does specialized operations on images. At least, that is what I usually do.

There are two good free C++ development environments. The standard tool for Windows PCs:

http://msdn.microsoft.com/vstudio/express/

The UNIX development evironment used to be fairly crude, but now there is a free knock-off of VC:

http://www.eclipse.org/

Posted by: gorelick Jun 16 2006, 03:45 AM

I haven't used them, but there are tools for this in ISIS: http://astrogeology.usgs.gov/Teams/Geomatics/pc.html
including a quick start guide: http://astrogeology.usgs.gov/Teams/Geomatics/photoclinometry/PC%2520userguide.pdf

Posted by: paxdan Jun 16 2006, 09:06 AM

Carnegie Mellon Researchers Teach Computers To Perceive Three Dimensions in 2-D Images

http://www.cmu.edu/PR/releases06/060613_3d.html

http://www.cs.cmu.edu/~dhoiem/projects/popup/index.html

Posted by: Bjorn Jonsson Jun 17 2006, 12:02 AM

Online (and useful) planetary DEMs are frustratingly hard to find, except for the Earth, Mars and Venus, even though DEMs have also been generated for Mercury, the Moon, the satellites of Jupiter and Saturn, Miranda and probably more. For the Moon, the best bet is probably to wait for LRO.

In contrast, there is a lot of theoretical material online, the most interesting stuff in the context of planets/satellites probably being the work of Randolph L. Kirk.

This one is very interesting:
http://astrogeology.usgs.gov/Projects/ISPRS/MEETINGS/Houston2003/abstracts/Kirk_isprs_mar03.pdf

And possibly also this one although I haven't read it yet:
http://people.csail.mit.edu/bkph/AIM/AIM-1105A-TEX.pdf

Using this search string on Google results in several interesting hits:
+"shape from shading" kirk

Posted by: Malmer Jun 18 2006, 11:19 PM

Using this search string on Google results in several interesting hits:
+"shape from shading" kirk
[/quote]


Searching on anything that starts with "shape from..." yields interesting reads. Shading, Stereo, Motion, Focus. try searching for "shape from X"

Shree k Nayar has a few really interesting papers on creating depth information from atmospheric scattering...



Mattias