Back to forums

Dark vertical streaks in stacked image

Equipment Artifacts Help Needed
15 replies•359 views
Richard Carande avatar
Richard Carande avatar
I've been struggling with trying to understand the source of these streaks I see in some recent stacked images.  I'm hoping someone here can offer an idea as to what the source is.  I'm hoping it's not the camera itself.  

My set up is a 300mm FRA Askar-pro OTA with a ASI2400MC-pro (full-frame) camera.  Control is from an ASIAIR pro.  I have used this set up before without this issue.  

I am capturing wide-field data from near LDN437 (the Gecko N.)  The image below is a (brightened for visibility) 133% zoom of an area near the nebula, however, these lines are throughout the image.  They run in the vertical direction, that is, along the short dimension of the camera sensor.   I'm using AstroPixelProcessor to stack and calibrate.  The image has been calibrated with flats and bias frames, no darks.  (The streaks are not visible in the flats.)  This image is a stack of 155 180-second exposures (7.45 hours) with no filter.  I'll note that when I've run with a dual-band filter (Ha/Oiii) and 300 second exposures I do not see these artifacts.  

I have swapped out the USB cable between the camera and the ASIAIR, thinking that might be the source, but the lines persist.  Also, I am dithering every three exposures -- which makes it hard to understand how these could persist if it's from the camera.  

In summary:  
-  Streaks are in broadband 180s exposure stack,
-  Streaks are not observed in Dual-band 300s exposure stack. 
-  Swapping out the USB cable to the camera did not change things
-  This is a recent observation; I've never noticed this before. 

Thoughts as to how to continue to troubleshoot this problem are welcome.  
Well Written Helpful Engaging
andrea tasselli avatar
andrea tasselli avatar
This stuff was (and is) not totally unexpected from a CCD sensor, faulty lines which do occur and are presistent. If they indeed are such they should also appear in dark frames and obviously in flats. So the question is: do they?
Richard Carande avatar
Richard Carande avatar
andrea tasselli:
This stuff was (and is) not totally unexpected from a CCD sensor, faulty lines which do occur and are presistent. If they indeed are such they should also appear in dark frames and obviously in flats. So the question is: do they?

They are not in the flats.  And again, since I'm dithering and this is a stack of 155 subs, wouldn't you expect them to be less distinct?
Richard Carande avatar
Richard Carande avatar
Richard Carande:
andrea tasselli:
This stuff was (and is) not totally unexpected from a CCD sensor, faulty lines which do occur and are presistent. If they indeed are such they should also appear in dark frames and obviously in flats. So the question is: do they?

They are not in the flats.  And again, since I'm dithering and this is a stack of 155 subs, wouldn't you expect them to be less distinct?

So I went back and looked at my master flat, and there are some vertical lines (!), which I didn't notice in individual flats. They are not as distinct as in the final stack.  But then shouldn't the flats have corrected the image during calibration.

Also, not sure if you are  just using the term more generically, but the 2400mc-pro is a CMOS, not CCD, sensor.
Steve Solon avatar
Steve Solon avatar
Hi Richard,
Those streaks are defective columns, something inherent in most CCD cameras as they age. If you use PixInsight, during the process CosmeticCorrection, you can create a list of all defective columns and rows, save that list, and apply it to all the subs when you run CC. I assume other processing programs would have something similar. My elderly QSI 683 has a defect list containing about 7 entries, but they're completely taken care of using the defect list. Hope this helps.
 - - Steve
Well Written Helpful Insightful Respectful Supportive
Richard Carande avatar
Richard Carande avatar
Steve Solon:
Hi Richard,
Those streaks are defective columns, something inherent in most CCD cameras as they age...

Thanks for the lead -- I don't use PI too much but will have a look.  Does your CCD statement also apply to CMOS cameras?
Well Written
andrea tasselli avatar
andrea tasselli avatar
Richard Carande:
So I went back and looked at my master flat, and there are some vertical lines (!), which I didn't notice in individual flats. They are not as distinct as in the final stack. But then shouldn't the flats have corrected the image during calibration.

Also, not sure if you are just using the term more generically, but the 2400mc-pro is a CMOS, not CCD, sensor.


*No, I'm referring to the actual CCDs not CMOS sensors. Which is all the more suprising since they (CMOS) are being read quite differently and thus single column defects are virtually unheard of, AFAIK. In either cases the removal isn't done with flats but with darks and, also, they are not 100% effective. One way around is to use the CosmeticCorrection procedure as mentioned above, but, again, they might not work that well if you have 2 defective columns one next to each other. If you can return the camera for a replacement then do so, otherwise the defects will increase as the sensor ages (tis was specially dramatic with the latest CCD sensors so CMOS could have a different ageing response).
Helpful
Richard Carande avatar
Richard Carande avatar
andrea tasselli:
Richard Carande:
So I went back and looked at my master flat, and there are some vertical lines (!), which I didn't notice in individual flats. They are not as distinct as in the final stack. But then shouldn't the flats have corrected the image during calibration.

Also, not sure if you are just using the term more generically, but the 2400mc-pro is a CMOS, not CCD, sensor.


If you can return the camera for a replacement then do so,

I bought the camera, used, in 2022.  So no returns for me at this point.  Thanks for the guidance.  I'll experiment with a few things as I can, to better characterize the issue.
Well Written Respectful
Steve Solon avatar
Steve Solon avatar
Richard Carande:
Steve Solon:
Hi Richard,
Those streaks are defective columns, something inherent in most CCD cameras as they age...

Thanks for the lead -- I don't use PI too much but will have a look.  Does your CCD statement also apply to CMOS cameras?

I have both a ZWO 2600 MC Pro and a 183MC Pro CMOS camera, and I have yet to see any defective columns or rows - perhaps I've just been lucky. The only defect in both (because they're older models) is amp glow, which is dealt with effectively using darks and properly calibrated flats. While PI does have just about every tool imaginable, I do use PhotoShop for final touch-ups - I have for years.
Well Written Concise
John Hayes avatar
John Hayes avatar

That is most likely a problem with the camera but it might not be the end of the world. You might try rotating the frame by 90 degrees and then use the CanonBandingReduction script in PixInsight. You can use an image container to perform the script on all of your subs before you stack. That should reduce the problem to an unnoticeable level.

John

Well Written Helpful Concise
Richard Carande avatar
Richard Carande avatar
UPDATE:  After replacing cables, and collecting all new calibration data (flats/darkflats/darks/bias) and not seeing any improvement, I went back through my recent use of the camera to see where i first noticed the issue.  The vertical line issue first appeared after a night in which it apparently rained at some point, but very lightly and for a short period.  The rig continued to operate through the 10 minute event, the skies cleared, and continued to collected apparently "good' data after that.  That's the first night I noticed the lines.  In subsequent nights the lines only got worse.  

I'll keep this camera out of rotation for a while, and hope it improves; I'm not very hopeful however.  Which is sad because this is my main camera for my high focal length rig (Edge-11) with its full frame and large pixels.  Also, it's expensive to replace.

I may try to program up a straightforward correction algorithm to see if I can detect and correct these anomalous lines.  I'm thinking if I simply average across the image and generate an image intensity profile, which should show dips due to the lines, I could apply its normalized inverse to the data.  In face, someone must have done this already.  I'll report back if anything positive comes of that.
Helpful Insightful Respectful Engaging
Carastro avatar
Carastro avatar
I wonder whether dithering would help overcome this problem.
Well Written
John Hayes avatar
John Hayes avatar

Richard Carande · Aug 22, 2025 at 09:27 PM


I may try to program up a straightforward correction algorithm to see if I can detect and correct these anomalous lines.  I'm thinking if I simply average across the image and generate an image intensity profile, which should show dips due to the lines, I could apply its normalized inverse to the data.  In face, someone must have done this already.  I'll report back if anything positive comes of that.

That’s basically what the CanonBandingReduction script does…

Tony Gondola avatar
Tony Gondola avatar

It might be interesting to open the camera up and have a look at the circuit board. Power flowing through a board that’s wet can quickly cause corrosion if it’s not comformally coated.

Well Written Insightful Concise
Willem Jan Drijfhout avatar
Willem Jan Drijfhout avatar

As others have mentioned, this is likely a camera issue. But there are a few things puzzling:
- Dithering does not seem to help
- In narrowband the effect does not show

As far as dithering goes, perhaps you could try a more aggressive dithering with the FRA300. Also, how does it look with the primary scope for this camera, the C11? The much smaller pixel-scale and appropriate dithering might reduce the problem.

Have you tried stacking without flats? Are you seeing the lines then? Just trying to rule out any option related to flats. Very short exposure times in flats can cause banding in the panel and some cameras have issues with sub-second shots. Your narrowband flats are probably longer duration (few secs?) negating both potential issues, which could explain why you don’t see it in NB. This is a long shot, but worth ruling out.

These were just a few thoughts, and likely no complete solutions. And even if no solution is found, a lot can be done in post-processing. The script that John mentioned can help, the linear defect option in CosmeticCorrection might work, and also there are LinearDefectDetection and the associated LinearPatternSubtraction scripts in PixInsight that might also help to reduce the problem. Always better than just giving up on an otherwise very lovely camera that should work very well on your C11.

Helpful Insightful Respectful Engaging Supportive
Richard Carande avatar
Richard Carande avatar
Willem Jan Drijfhout:
As others have mentioned, this is likely a camera issue. But there are a few things puzzling:
- Dithering does not seem to help
- In narrowband the effect does not show

As far as dithering goes, perhaps you could try a more aggressive dithering with the FRA300. Also, how does it look with the primary scope for this camera, the C11? The much smaller pixel-scale and appropriate dithering might reduce the problem.

Have you tried stacking without flats? Are you seeing the lines then? Just trying to rule out any option related to flats. Very short exposure times in flats can cause banding in the panel and some cameras have issues with sub-second shots. Your narrowband flats are probably longer duration (few secs?) negating both potential issues, which could explain why you don’t see it in NB. This is a long shot, but worth ruling out.

These were just a few thoughts, and likely no complete solutions. And even if no solution is found, a lot can be done in post-processing. The script that John mentioned can help, the linear defect option in CosmeticCorrection might work, and also there are LinearDefectDetection and the associated LinearPatternSubtraction scripts in PixInsight that might also help to reduce the problem. Always better than just giving up on an otherwise very lovely camera that should work very well on your C11.

Thanks for these ideas.  I have tried to stack without using any flats, and while the details of the artifacts changed, the problem was still there.  In several years of using this camera, I've never experienced this issue before.  I haven't had a chance to test it out on my HD11, but will as soon as I can.
Well Written