Howdy,

I'm new to CCD Stack and so far am loving it, EXCEPT I cannot get my images to perfectly allign using the star snap method.

The stack is very close, but there is consistantly a couple of images that simply will not allign perfectly. I've tried on various data, same result.

I typically choose a star close to each corner, then one last in the center.

Is there a trick to getting good registration of images? I try correcting via manual input, but it is like herding cats.

Thanks for any help!

JOhn

Hi John,

1) Be sure to avoid saturated stars or close-in (crowded) stars. Stars that are affected by PSF distortions may produce inconstant centroids.

2) Maybe the scope's field is distorted (e.g. "pincushion") and makes it impossible to fully and accurately register images with non-trivial offsets. This is more common than you might think. Select reference stars nearest the center of the frames (you only need 2). BTW, we are contemplating a future version of CCDStack/CCDIs that warps (de-warps) frames to deal with this issue.

3) Maybe the frames are correctly aligned (do they look OK in the red overlay?) but you are using Nearest Neighbor to resample and when you blink the stack you become alarmed by stars that wiggle around? This is an artifact of Nearest Neighbor because it can only register via integer pixel offsets. The effect is most noticeable for undersampled images and/or rotational alignments.

If you would like further assistance then please send me 2 frames that illustrate the problem (stan_ccd at yahoo dot com).

Regards,
Stan

Hi Stan,

Thanks for your reply. I have tried the various techniques you mention (diff star selections, nearest neighbor, bicubic b spline) but still experience mis-registration. Most frames shift a bit, but its mostly 2 particular images that exhibit gross problems. Blinking the results show the images shifting about way too much.

I jumped over to Maxim to try. Using the Automatic routine kicked out the same 2 offending frames (I believe they are slightly soft focus compared to the others, but only slight) BUT I managed to get perfect alignment using Maxim's manual 2-star method.

I went back to CCD STack and tried a simple "2-star" alignment there, but still the same problem.

By the way, I was using a FSQ 106EDX and STL-11000M.

Any other thoughts?

Best,

JOhn

please contact me (stan_ccd at yahoo dot com) to arrange to send the images so that I can examine the problem.

Thanks,
Stan

Just sent you an email -- thanks!

JOhn