| Re: matching coordinates [message #55055] |
Mon, 30 July 2007 17:20 |
JD Smith
Messages: 850
Registered: December 1999
|
Senior Member |
|
|
On Mon, 30 Jul 2007 11:12:23 +0000, jradavenport wrote:
>
> % HISTOGRAM: Illegal binsize or max/min.
>
> This is for the line:
> ;; Dual HISTOGRAM method, loop by repeat count in bins
> h2 = histogram(h[b],MIN=1,REVERSE_INDICES=ri2)
>
> Anybody have ideas who knows about Histogram?
That is due to an array of all zeroes being passed to HISTOGRAM, with
MIN=1. It seems that *none* of your bins had a match in it, which is
very likely an indication that you need a larger search radius.
I've rolled the original algorithm into the function MATCH_2D,
available here:
http://turtle.as.arizona.edu/idl/match_2d.pro
This version deals gracefully with the case of no search points in any
target point bin, and also returns -1 in all cases where no point was
found within the search radius (vs. returning a
close-but-not-necessarily-the-closest point).
In principle one could also write MATCH_ND, but the sensitivity to a
poorly chosen search_radius would become quite extreme in dimensions
higher than 2, as would the need to search 2^d adjacent bins.
JD
|
|
|
|
| Re: matching coordinates [message #55062 is a reply to message #55055] |
Mon, 30 July 2007 07:58  |
Conor
Messages: 138
Registered: February 2007
|
Senior Member |
|
|
On Jul 30, 7:12 am, jradavenp...@gmail.com wrote:
> On Jul 30, 4:36 am, jradavenp...@gmail.com wrote:
>
>
>
>> On Jul 29, 7:53 am, David Fanning <n...@dfanning.com> wrote:
>
>>> jradavenp...@gmail.com writes:
>>>> Hello, I'm a long time lurker of this group, and finally have a
>>>> problem I can't solve efficiently enough! I have coordinates for stars
>>>> (RA and DEC for us astronomers out there), from two sources. i.e.
>>>> (x1,y1) and (x2,y2). These lists are huge, like 150k for one and
>>>> 5million in the other (a lot of stars!). I need to find the matches
>>>> from these catalogs within a tolerance. I've tried a few programs
>>>> I've found online (close_match_radec for instance) and they have not
>>>> given me results I believe (multiple matches and such). Here is some
>>>> horrible code I wrote just now which solves the problem VERY slowly:
>
>>> Here is an article that might help:
>
>>> http://www.dfanning.com/code_tips/matchlists.html
>
>>> For background, you might want to read the precursor
>>> article, and one of my personal favorites:
>
>>> http://www.dfanning.com/code_tips/slowloops.html
>
>>> Cheers,
>
>>> David
>>> --
>>> David Fanning, Ph.D.
>>> Fanning Software Consulting, Inc.
>>> Coyote's Guide to IDL Programming:http://www.dfanning.com/
>>> Sepore ma de ni thui. ("Perhaps thou speakest truth.")
>
>> Thanks David, Big fan of your site and of yours and JD Smith's code...
>> I guess it really is time to learn that Histogram function (something
>> I've put off for over a year!)
>
>> I'm turning that example code into a workable program, but am confused
>> on where the results are. Which Variable do I want to return with the
>> answers?
>
> Actually, scratch that. I believe it's the min_pos and min_dist
> variables. The only other problem I'm having is it's telling me in
> line 40 (including a program begin line)
>
> % HISTOGRAM: Illegal binsize or max/min.
>
> This is for the line:
> ;; Dual HISTOGRAM method, loop by repeat count in bins
> h2 = histogram(h[b],MIN=1,REVERSE_INDICES=ri2)
>
> Anybody have ideas who knows about Histogram?
As an alternate solution, here's what I use:
astro.ufl.edu/~cmancone/pros/qfind.pro
I wrote this code as a result of the discussion here:
http://groups.google.com/group/comp.lang.idl-pvwave/browse_t hread/thread/629cbb2a852c5371/6ada6d1659bc55a7?hl=en#6ada6d1 659bc55a7
You would call:
result = qfind(x1,y1,x2,y2,posshift=tolerance)
'result' would be a 2xn array where n is the total number of matched
stars. result is essentially the equivelent of two where functions.
So for instance:
abs( x1[result[0,*]] - x2[result[1,*]] ) is always less than tolerance
x1,y1,x2,y2 all need to be row vectors. If you pass them as column
vectors, they will automatically be transposed (which, now that I
think about it, might be bad if the rest of your program expects
column vectors). I'm not sure how it compares to the other for speed
(probably worse), but I'd be curious.
|
|
|
|
| Re: matching coordinates [message #55065 is a reply to message #55062] |
Mon, 30 July 2007 04:12 |
jradavenport
Messages: 6
Registered: July 2007
|
Junior Member |
|
|
On Jul 30, 4:36 am, jradavenp...@gmail.com wrote:
> On Jul 29, 7:53 am, David Fanning <n...@dfanning.com> wrote:
>
>
>
>> jradavenp...@gmail.com writes:
>>> Hello, I'm a long time lurker of this group, and finally have a
>>> problem I can't solve efficiently enough! I have coordinates for stars
>>> (RA and DEC for us astronomers out there), from two sources. i.e.
>>> (x1,y1) and (x2,y2). These lists are huge, like 150k for one and
>>> 5million in the other (a lot of stars!). I need to find the matches
>>> from these catalogs within a tolerance. I've tried a few programs
>>> I've found online (close_match_radec for instance) and they have not
>>> given me results I believe (multiple matches and such). Here is some
>>> horrible code I wrote just now which solves the problem VERY slowly:
>
>> Here is an article that might help:
>
>> http://www.dfanning.com/code_tips/matchlists.html
>
>> For background, you might want to read the precursor
>> article, and one of my personal favorites:
>
>> http://www.dfanning.com/code_tips/slowloops.html
>
>> Cheers,
>
>> David
>> --
>> David Fanning, Ph.D.
>> Fanning Software Consulting, Inc.
>> Coyote's Guide to IDL Programming:http://www.dfanning.com/
>> Sepore ma de ni thui. ("Perhaps thou speakest truth.")
>
> Thanks David, Big fan of your site and of yours and JD Smith's code...
> I guess it really is time to learn that Histogram function (something
> I've put off for over a year!)
>
> I'm turning that example code into a workable program, but am confused
> on where the results are. Which Variable do I want to return with the
> answers?
Actually, scratch that. I believe it's the min_pos and min_dist
variables. The only other problem I'm having is it's telling me in
line 40 (including a program begin line)
% HISTOGRAM: Illegal binsize or max/min.
This is for the line:
;; Dual HISTOGRAM method, loop by repeat count in bins
h2 = histogram(h[b],MIN=1,REVERSE_INDICES=ri2)
Anybody have ideas who knows about Histogram?
|
|
|
|
| Re: matching coordinates [message #55066 is a reply to message #55065] |
Mon, 30 July 2007 03:36 |
jradavenport
Messages: 6
Registered: July 2007
|
Junior Member |
|
|
On Jul 29, 7:53 am, David Fanning <n...@dfanning.com> wrote:
> jradavenp...@gmail.com writes:
>> Hello, I'm a long time lurker of this group, and finally have a
>> problem I can't solve efficiently enough! I have coordinates for stars
>> (RA and DEC for us astronomers out there), from two sources. i.e.
>> (x1,y1) and (x2,y2). These lists are huge, like 150k for one and
>> 5million in the other (a lot of stars!). I need to find the matches
>> from these catalogs within a tolerance. I've tried a few programs
>> I've found online (close_match_radec for instance) and they have not
>> given me results I believe (multiple matches and such). Here is some
>> horrible code I wrote just now which solves the problem VERY slowly:
>
> Here is an article that might help:
>
> http://www.dfanning.com/code_tips/matchlists.html
>
> For background, you might want to read the precursor
> article, and one of my personal favorites:
>
> http://www.dfanning.com/code_tips/slowloops.html
>
> Cheers,
>
> David
> --
> David Fanning, Ph.D.
> Fanning Software Consulting, Inc.
> Coyote's Guide to IDL Programming:http://www.dfanning.com/
> Sepore ma de ni thui. ("Perhaps thou speakest truth.")
Thanks David, Big fan of your site and of yours and JD Smith's code...
I guess it really is time to learn that Histogram function (something
I've put off for over a year!)
I'm turning that example code into a workable program, but am confused
on where the results are. Which Variable do I want to return with the
answers?
|
|
|
|
|
|
|
|
| Re: matching coordinates [message #55075 is a reply to message #55071] |
Sun, 29 July 2007 06:53 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
jradavenport@gmail.com writes:
> Hello, I'm a long time lurker of this group, and finally have a
> problem I can't solve efficiently enough! I have coordinates for stars
> (RA and DEC for us astronomers out there), from two sources. i.e.
> (x1,y1) and (x2,y2). These lists are huge, like 150k for one and
> 5million in the other (a lot of stars!). I need to find the matches
> from these catalogs within a tolerance. I've tried a few programs
> I've found online (close_match_radec for instance) and they have not
> given me results I believe (multiple matches and such). Here is some
> horrible code I wrote just now which solves the problem VERY slowly:
Here is an article that might help:
http://www.dfanning.com/code_tips/matchlists.html
For background, you might want to read the precursor
article, and one of my personal favorites:
http://www.dfanning.com/code_tips/slowloops.html
Cheers,
David
--
David Fanning, Ph.D.
Fanning Software Consulting, Inc.
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Sepore ma de ni thui. ("Perhaps thou speakest truth.")
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
