[AstroPy] Python version of IDL Astron library

Ray S rays at blue-cove.com
Tue Sep 7 16:07:59 CDT 2004


<x-flowed>Speaking of migrating from IDL to Python, I would like to see some 
discussion of the issues/problems others have uncovered in their ports.

I have been trying to port Hongjie Xie's IDL/ENVI Implementation of the FFT 
Based Algorithm for Automatic Image Registration
http://www.nmt.edu/%7Ehjxie/xie-paper.pdf PDF
ftp://ftp.iamg.org/VOL29/v29-08-10.zip sources
which calculates shift, scale and rotation for 2d arrays, to 
Python/numarray/Numeric and running into some subtle problems.

As small examples, the IDL ROT() requires some additional steps to emulate, 
and the MAX(array, position) as well.
I'm not at all familiar with the correspondence of the FFT libraries... 
even the simple shift-only routine of mine
http://rjs.org/astro/1004x/Python/register/shift_idl.py
produces different (incorrect) x shift values than Xie's IDL.

My current code
http://rjs.org/astro/1004x/Python/register/
conforms to Xie's flow as much as possible for now, but should and will be 
re-factored when functional. His LogPolar function, to convert rectangular 
coordinate arrays to polar, is particularly slow.

Dr. Xie said that he is available for specific assistance, but Python is 
not part of his repertoire (although C is).

I was recently informed that IDL has a usable trial, so I am trying it out 
this week, but if anyone has an interest in this general array 
registration, please let me know... I think that the algorithm would make a 
nice standalone module, or perhaps something for nd_image.


At 03:09 PM 9/7/2004 -0400, Paul Barrett wrote:

>I have refrained from listing the procedures in each category, since there 
>are a large number of them.  Instead, I will add a page to the AstroPy web 
>site in the next few days, if there is sufficient interest in the project 
>by the AstroPy community.  I do not expect developers to devote are large 
>percentage of their time to this project, unless they want to of course, 
>but to contribute two or three procedures that are of interest to them 
>over the next year.

>Any takers?

I am interested in developing registration methods, at least. I might be 
able to enlist the help of some SANPYT members as well; Dr. Chris Seberino 
was at SciPy 2004, and I'd only heard some tidbits from him as yet.


_________________________________________________
AstroPy mailing list     -      astropy at stsci.edu
http://www.astro.washington.edu/owen/AstroPy.html
</x-flowed>


More information about the AstroPy mailing list