Dark Column

[viktorbravo]

I am having trouble getting rid of a dark column in my images.  My images are dithered and of course calibrated with a master dark.  I have not used cosmetic correction as it seems to reject good low signal pixels from my images so I am trying to get rid of it during the integration process using windsorized sigma clipping.  The rejection map shows that it was rejected, however its still apparent in the stretched image.  I have used different levels of sigma with the same result, shown in the rejection map, but not totally rejected in the final image. 

Other software gets rid of it, but I am trying to use PI for my entire process workflow.

Any ideas on what I am doing wrong?

See the upper left portion of the image:  http://crashmaster.smugmug.com/Other/Astroimages/19347496_tcX7Pn#!i=1932220962&k=rjxMSdc&lb=1&s=O

[viktorbravo]

Nadie sabe nada?   

[Geoff]

Creating a defect map may help.  See this thread http://pixinsight.com/forum/index.php?topic=4218.msg29598#msg29598
Geoff

[Juan Conejero]

I have not used cosmetic correction as it seems to reject good low signal pixels from my images

Why? Have you tried it in "defect list" mode? See the attached screenshot. Note that this is a quick test with your JPEG image - you should apply this process to each of your raw frames. You can do this very easily with the BatchPreprocessing script.

I am trying to get rid of it during the integration process using windsorized sigma clipping.  The rejection map shows that it was rejected, however its still apparent in the stretched image.

That's strange. In general, a bad column is a well defined structure that can be rejected easily without uncertainty. If the low rejection map shows it as a set of parallel vertical lines (one line for each integrated image, since you are dithering), I don't understand why it hasn't been rejected. If the default "sigma low" value doesn't work, you should be able to get rid of these bad columns by simply lowering this parameter a bit. Or is this column defect more complex than a single column of abnormally dark pixels?

Anyway, the correct way to fix these defects is by applying CosmeticCorrection to each individual raw frame.

[viktorbravo]

Thanks for the tips Juan I'll give the defect map a try a see what happens.  Its curious that the integration rejected pixel map shows it being removed.  I'll try CC and see if I can get a decent result.

Muchas gracias.

[viktorbravo]

I was able to produce a defect map that worked very well, then used the auto detect with sigmas of 2.0 to get rid of the remaining hot and cold pixels.  Windsorized Sigma Clipping in Image Integration took care of the rest of the cosmic ray bird poop very well.  Now I think I have a handle on this cosmetic correction thing.  Next I need to figure out how to build a defect map using a master dark.  Baby steps.

This is a 50% res image:

http://crashmaster.smugmug.com/Other/Astroimages/19347496_tcX7Pn#!i=1932220962&k=rjxMSdc&lb=1&s=O

Now I just need some clear skies.  Is there a setting for that in PixInsight?   

Thanks again for your help Juan.

[cardiofuse]

The CC works great for the two "dark" columns mysteriously introduced after calibration of my images of NGC7293