iraf.net

Hi, I'm using imalign in order to align images for blinking and
perform a moving object deteccion.
When using imalign, on images of fwhm ob about 12 pix, I get:imalign @object_list 2006hh123_r_420s_a1.imh archivo @object_list#Coords Image X-center Err Y-center Err Num
2006hh123_r_420s_b3.imh 423.551 (0.026) 863.978 (0.030) 1
2006hh123_r_420s_b3.imh 423.551 (0.026) 863.978 (0.030) 2
2006hh123_r_420s_b3.imh 224.164 (0.062) 824.299 (0.072) 3
2006hh123_r_420s_b3.imh 818.428 (0.034) 646.793 (0.040) 4
2006hh123_r_420s_b3.imh 1127.395 (0.021) 828.704 (0.024) 5
2006hh123_r_420s_b3.imh 1038.887 (0.078) 703.480 (0.095) 6
2006hh123_r_420s_b3.imh 647.089 (0.082) 618.886 (0.096) 7Warning: failed to converge near (449,882)
Warning: Problem with coarse centering
ERROR: No shifts were calculated.And that's it. What I'm using is a box of 45 and a bigbox of 51. Interpolant spline3. Archivo is the file with stars centers of the reference image.
I'd used several different values, e.g. 7 and 11 for box and bigbox respectively, with no good results.
Any suggestion?
If this works, could I display the shifted images and get a nice blinking with "fixed" stars?
best regards
sebastian

I think the problem is in the size of your boxes: For a fwhm of 12 pix a centering box/bigbox of 7/11 is much too small, and for 45/51 it's too big and may include neighboring stars in the centroid causing the failure. Choose a size slightly bigger than the fwhm and be sure your reference coords are correct.Otherwise, if the images already have a WCS you might just look at WREGISTER as a simpler way to align the images. Other tasks in the IMMATCH package can also help.-Mike

Hello,
I am trying to align 3 images in order to combine them using imcombine. I've been using imalign to do this. When I run imalign on images
And8K004Zfinobj_200.fits And8K004Zfinobj_300.fits And8K004Zfinobj_400.fits
the process seems to run fine until it reaches the below warning. "failed to converge near (1780,23)". I've tried changing my boxsize and bigbox size from 4/6 to 9/12, 10/14, 26/30 ect and it appears the higher I go the faster I get the error.
Here is the input I have set for one of my calls of imalign. Where the .cat file is a file with xy coordinates for the first reference image and 8k001 is the offsets calculated for each image.imalign ('@zero',"And8K004Zfinobj_200.fits", "And8K004Zfinobj_201.cat", '@fzero', shifts="8K001_1off.ls", boxsize=4, bigbox=6, negative=no, background=INDEF,lower=INDEF, upper=INDEF, niterate=3, tolerance=0, maxshift=INDEF, shiftimage=yes, interp_type='linear', boundary_typ='nearest', constant=0., trimimages=yes, verbose=yes, list='')AND HERE IS THE OUTPUT when it is ran:
And8K004Zfinobj_200.fits 239.705 (0.149) 3143.785 (0.175) 177
And8K004Zfinobj_200.fits 241.423 (0.096) 3995.426 (0.143) 178
And8K004Zfinobj_200.fits 239.568 (0.128) 4020.707 (0.154) 179
Warning: failed to converge near (240,3821)
And8K004Zfinobj_200.fits 1840.595 (0.139) 4048.468 (0.145) 181
And8K004Zfinobj_200.fits 142.564 (0.195) 3957.516 (0.135) 182
And8K004Zfinobj_200.fits 1820.840 (0.178) 3970.929 (0.138) 183
And8K004Zfinobj_200.fits 1649.256 (0.182) 3848.009 (0.163) 184Warning: failed to converge near (1780,23)
Warning: Problem with coarse centering
ERROR: No shifts were calculated.
called as: `cl (gcur=)'I'm not sure what my ob FWHM is, but it seems to not be a major issue because in previous months data I used boxsize=5 bigbox=9 and it ran fine. It appears to be a consistent problem with this months data though.
Any suggestions?
Thank you,
Shelly

The box/bigbox parameters are used to set the size of the centering box for a specific object, not the box size in which the shifted object is expected to be found. Think of it this way: A box size of 512x512 will certainly contain your object, it will also likely contain hundreds of other objects and doing a simple marginal centroid of that box may not ever converge.The 'warning: Problem with coarse centering' is coming from the IMCENTROID task called by the IMALIGN script and is the thing that starts the real problems. This message simply means the centering didn't converge within the number of iterations set by the 'niterate' parameter. You can increase the niterate value, but you should also look at the image to see what's at those coords and if it's a close pair of stars simply delete it from the coords list. Without seeing the image, I can't tell you what an appropriate value for box/bigbox should be.-Mike

Mike,
Thank you for your response. It helped! I appreciate you taking the time to explain each area for me. I was a bit confused by this at first, but your explanation helped. I was able to delete the one bad region from the list and continue to use imcombine without a problem.
Thanks again!
Shelly

I too had a similar problem with imalign. I was giving it all the appropriate inputs and it would continuously report the "coarse centering" error. After two days of struggling with it, I figured out that I was improperly including an "@" symbol with my shifts file. This apparently causes IRAF to ignore the (shifts) input altogether and assume that it will have to compute the offsets itself.So, always mind your "@" symbols when using IRAF.