PixInsight Forum (historical)
PixInsight => General => Topic started by: John_Gill on 2018 May 22 02:38:00
-
Hi,
Using a APM APO 107-700 refractor and Canon 600d DSLR. I have 3 hours of data from night 1 and a further 1 hour from night 2. Each batch was processed separately and they look good.
I then took the "Debayered" images from each night and combined them with StarAlignment, ImageIntegration, LocalNormalization, DrizzleIntegration, DynamicCrop and PhotometricColorCalibration. The result was great except about 40 stars scattered across the image have black artifacts in their centers. These are not the brightest or big fat saturated stars. I tried doing a " Repair HSV Separation" and then ChannelCombination, but that didn't help. I guess I can use CloneStamp but would prefer to fix the issue.
Any ideas would be appreciated.
Look up
John
-
are the black artifacts right on the centers? sometimes artifacts at the edges of stars can be caused by the default Lancosz-3 interpolation method that StarAlignment uses. it might be worth playing with the clamping threshhold, or trying bicubic spline instead of Auto.
rob
-
Thanks Rob, I will give it a try. I have attached a small preview of the strange stars .....
-
ok, no that's something else. all 3 channels are like this? are all the calibrated subs like this? or just the integration?
rob
-
John
If you can't find a good or proper solution here are a couple of possible ways of making a mask for some of the offending stars. This isn't perfect and I would not suggest this as a first thing to do but it might come in handy.
Mike
-
Hi,
The individual subs are perfect. I have processed each nights batch separately without issue. I then combined all the images, without drizzle and again no problem. The artefacts only appear after doing the DrizzleIntegration. I did a ChannelExtraction in RGB and in
CIE L* a*b and the artefacts exist in RGB and in the L channel.
Look up
John
-
interesting. are you using pixel rejection in ImageIntegration or DrizzleIntegration? for fun you might try turning that off in each place in turn to see what happens.
rob
-
Hey Rob,
I did a DrizzleIntegration with Pixel Rejection turned off and no more funny artefacts. Thank you for your help and suggestions.
Look up
John
-
what is weird about this is that i vaguely remember that there was a bug similar to this regarding drizzle integration and pixel rejection, but i thought it had been fixed...
rob
-
rob
I think this is the bug you might be thinking about.
https://pixinsight.com/forum/index.php?topic=10362.msg65316#msg65316 (https://pixinsight.com/forum/index.php?topic=10362.msg65316#msg65316)
Mike
-
yes, that is it... John, is it possible you are using a downrev version of PI and/or updates have not been applied in a while?
rob
-
Hi,
My version is 01/08.05.1353 Ripley (x64) which I think is the latest version. There have not been any updates for a few months. This is running on a Windows 10 version 1709 laptop.
Look up
John
-
I have not used the drizzle feature.
Can you use an roi over the stars during integration and check how the percentile/sigma/linear/high settings are affecting your stars at their given setting? Sometimes I have issues with hotpixels not being rejected if the settings are to high.
Additionally, the only time I can remember coring out my stars was when I used cosmetic correction too aggressively, I realized after integrating that alot of my fainter undersampled stars had been hollowed out, while in my sample window I had not created any damage.