Question about image integration and rejection algorithm

[gvanhau]

Hello

Yesterday evening i was integrating a stack of 38 images usign sigma clipping rejection algorithm.
I was surprised that I got a big stain (like a light flare) all over the resulting image.
Revising the integration set, I got one single image with simmilar flare.

Im curious if this this spot, shouldn't have been rejected during the integration process?
It was to late to test other rejection algorithms, simpy deselected the "bad" image and reprocessed.

Regards
Geert

[Alejandro Tombolini]

Hi Geert,
I usually use Winsorized sigma clipping and there are no problem in rejection.
Sigma Clipping should have also rejected it, but may be is something strange that is not recognized. It would be a good experiment if you can check other algoritms

Saludos. Alejandro.

[gvanhau]

Thanks Alejandro
I will do this eveneing.

Regards
Geert

[gvanhau]

Hello
I tested all algorithms, and all of them got in trouble with the image included. (at least with default parameters).

I attached the resulting integration on the left and the bad image on the right.

regards
Geert

[sreilly]

My first suggestion is to cull through all the individual images and see if this is in all the images. I only use the best of the images I have. There may be a few that are ruining the image. As far a which method to intergrate, I use the following that Juan posted some time back:

% Clipping for 2-6 images
Sigma Clipping for 8-10 images
Winsorized Sigma Clipping for 15+ images
Averaged Sigma Clipping for 10+ images
Linear Fit Clipping for 25+ images

I always have Clip High/Low pixels checked as well as clip High/Low range

HTH

[gvanhau]

Thanks Steve.

I already inspected all images, and there was only one "bad" image.

My concern was  "why did only one image have such a bad influence on the final result,
1 over 38 should have been rejected with the rejection algorithm..."

If I disable this image, integration result is ok.

Regards
Geert