IDL / GDL for image processing Options

[scalbers]

Thought I'd mention that IDL is a nice language for doing image processing. It does require a fair amount of programming though it gives one a lot of flexibility. IDL licenses can be costly, however there is a new freeware version of IDL called GDL. There are a number of library procedures that come with the package, and more can be added by either writing your own or by scouring the web. I use IDL quite a bit in a Linux environment.

Here is a link to download the GDL software: http://gnudatalanguage.sourceforge.net/

This page has binaries for the Mac and source code for Linux. For Windows, one could use Cygwin to run Linux, though I have yet to get the GDL source code to fully compile in that environment.

[john_s]

IDL is the language used for image processing by most astronomers that I know (I personally use it for almost everything). It's extremely powerful, and extensive libraries of IDL routines are available (e.g., the IDL Astronomy Library). There's an interesting cultural divide in the community, though- planetary geologists don't seem to use IDL much, preferring more specialized packages like ISIS.

[NGC3314]

Yes, IDL is very popular in the astronomical community - for my optical/UV work, what IDL doesn't handle probably has an IRAF task. My quick check of the GDL free counterpart has been that the numerical routines are there, what it lacks is the ability to make really nice-looking plots (like EPS files with journal-quality fonts). I can see that a lot of planetary work starts with projection to/from spheroidal surfaces, which IDL has pretty much built in through its map projection native functions. IDL and IRAF are both rich in was to directly manipulate pixel data. One quick example - this might have looked just as good in Photoshop, but after an unreasonable amount of time I manage to mosaic 14 images over an exposure range of 2000 to get this lunar-occultation depiction (click on image to magnify).


Of course, most of the time we're interested in numerical output and error bars rather than aesthetic quality, but there are times it's simply irresistible. (Hi, John, last I knew you were in Flagstaff...)

[edstrick]

I suspect another reason that planetary scientists tend to not use IDL is that digital image processing was substantially invented by planetary scientists <or their supporting programmers>, starting at JPL. They invented the necessary tools first, and kept adding tools to JPL's VICAR, Flagstaff's IPSY and later ISIS software packages that were essentially necessary to handle mission specific data.

As a result, grad students would train on the tools available in-house and continue to use them as they progressed in the planetary science community.

[scalbers]

IDL/GDL can be used to reproject satellite/spacecraft images. So far I've found it easier to do this with my home-grown IDL procedures rather than the supplied library routines. I could try to describe either approach if I can organize things enough. To start with here are a couple of links to some of the "standard" ways of doing this:

http://www.msi.umn.edu/software/idl/tutori...ges%20to%20Maps

http://astro.uni-tuebingen.de/software/idl...o.html#IDLLAB13

[rlorenz]

I suspect another reason that planetary scientists tend to not use IDL is that digital image processing was substantially invented by planetary scientists <or their supporting programmers>, starting at JPL. They invented the necessary tools first, and kept adding tools to JPL's VICAR, Flagstaff's IPSY and later ISIS software packages that were essentially necessary to handle mission specific data.

John_S said planetary GEOLOGISTS. People who look at pictures.

There are whole sections of the planetary science community who do not, and most of them work with
IDL. Because IDL is a programming language that lets you build a model that you specify to whatever
algorithmic detail you want, and then plot the data (or generate a synthetic image to compare with real
data) it is very powerful and appealing. You can't build a radiative transfer model of an atmosphere with
ISIS, but you can with IDL. You can project images in both.

A lot of the geology types are moving to GIS systems (sometimes using ISIS to do the hard math
of reprojection to make the GIS shapefiles). I still perform GIS-like functions in my own clunky
IDL code.

The fact that I use IDL aside, I detest RSI..

[scalbers]

As it turns out I'm continuing to attempt an installation of GDL in Cygwin under Windows XP. Has anyone succeeded in doing this?

Thanks,

Steve

[emtilt]

Does anyone know how to add the IDL Astronomy Users library to GDL? I installed GDL on an xubuntu system and downloaded the library, but can't figure out how to add the commands.