Author Topic: Difficulty removing bad column  (Read 2398 times)

Offline astrodoc71

  • PixInsight Enthusiast
  • **
  • Posts: 92
    • View Profile
Difficulty removing bad column
« on: 2017 April 04 06:48:13 »
Strange issue with bad column removal. I noticed it in the final image. Not glaringly obvious but it was there. So I went back and sure enough it was very subtle but present in the supposedly calibrated images. I was using batch preprocessing. Checked all of the dark masters and the bad column is most definitely present so not sure why it is not being completely removed. However, when I checked "optimize dark frames" it did improve. I haven't checked integration yet with the optimized images but I can still see the column very very faintly. Just wondering what else I can try
Thanks
Dave

Offline aworonow

  • PixInsight Addict
  • ***
  • Posts: 256
    • View Profile
    • Faint Light Photography
Re: Difficulty removing bad column
« Reply #1 on: 2017 April 04 07:31:46 »
I've used TGVInPaint with some success. For cosmetic reasons, you may need to subsequently apply PixelMath to add some level of noise to the infilled pixels.
Alex

Offline chris.bailey

  • PixInsight Addict
  • ***
  • Posts: 235
    • View Profile
Re: Difficulty removing bad column
« Reply #2 on: 2017 April 04 08:38:29 »
I use the Defect List section of Cosmetic Correction. Use the Limit settings if the column defect is only partial. My H16 has a few and it works very well.

Chris

Offline Duncan

  • Member
  • *
  • Posts: 53
    • View Profile
Re: Difficulty removing bad column
« Reply #3 on: 2017 April 04 10:50:14 »
I also use the defect list in cosmetic correction; works great.

I've found that the easiest way to create a defect list is to open up a flat frame with strong stretching applied to show where the bad pixels/columns are.

If you haven't used this yet, the cosmetic correction tool automatically sets the image coordinates when you click on the column or pixel in the image. Makes it a breeze to make the list, which you can save for future use. I typically create these right after building calibration masters.

Offline astrodoc71

  • PixInsight Enthusiast
  • **
  • Posts: 92
    • View Profile
Re: Difficulty removing bad column
« Reply #4 on: 2017 April 08 14:21:07 »
So I used the defect list option in cosmetic correction and like you say it's pretty straightforward particularly as my sensor has really just 1 bad column. The individual cosmetically corrected frames to me look better but the final integrated result is still the same??? So not sure what to do. I am pretty certain I am applying it correctly. I tried using CC as a new instance during batch preprocessing and also applying CC after calibration was complete. No difference. I created the defect list by opening the cosmetic correction process, clicking on the "defect list" box at the bottom, opening a raw uncalibrated frame, clicked on "column" at the bottom of the defect list window and then clicked on a single pixel in the bad column of the raw frame. The defect then appears in the list with what I believe is the X value of the column. I save as a text file and then reload during the cosmetic correction process. Maybe this is wrong I am hoping? Not sure why there would be zero difference in the end result. I used the same integration with both (windsorized sigma) I have the 2 files attached. PI2 is cosmetically corrected WITHOUT defect map included. PI3 is WITH defect map included for supposed correction
« Last Edit: 2017 April 08 18:14:42 by astrodoc71 »

Offline CharlesW

  • PixInsight Enthusiast
  • **
  • Posts: 87
    • View Profile
Re: Difficulty removing bad column
« Reply #5 on: 2017 April 08 20:59:06 »
I think it needs asking, how many exposures is this, are you dithering, and for how many pixels?

Offline chris.bailey

  • PixInsight Addict
  • ***
  • Posts: 235
    • View Profile
Re: Difficulty removing bad column
« Reply #6 on: 2017 April 08 23:00:23 »
Worth double checking you have picked the right column by zooming right in.

I know the Batch Processing Script makes life a little easier but with problems like this it is worth going through the pre-processing steps one at a time so you can inspect at each stage

Image Calibration
Cosmetic Correction
Star Alignment
Image Integration

Offline astrodoc71

  • PixInsight Enthusiast
  • **
  • Posts: 92
    • View Profile
Re: Difficulty removing bad column
« Reply #7 on: 2017 April 09 05:29:57 »
Zoomed in to click on a single pixel in the column. I did it again just to be sure and the number was exactly the same. Dithering is 1.1 pixel. 29 frames in this stack
Thanks
Dave

Offline astrodoc71

  • PixInsight Enthusiast
  • **
  • Posts: 92
    • View Profile
Re: Difficulty removing bad column
« Reply #8 on: 2017 April 09 10:15:12 »
Ok I believe I am onto it now. I just discovered something! When I tried making a column mask, the line appeared in the wrong place! I used the Y value from the defect list and clearly it was not in the same location as the linear defect I am assuming is a "bad column". It is close to where the bad column is but it appears to be something different! The bad column is located at 1112. This has clearly been removed by calibration. The artifact in question is actually at 1158 or so. I have attached a raw uncalibrated image showing the bad column (dashed arrow) and the linear artifact which appears in the final image which is very faintly visible in the raw image. You can see they are different!
So my question is, what is it? How do I remove it? Could it be another sensor defect and should I include it in the defect map or try to eliminate it with post processing?
Thanks
Dave

Offline chris.bailey

  • PixInsight Addict
  • ***
  • Posts: 235
    • View Profile
Re: Difficulty removing bad column
« Reply #9 on: 2017 April 10 00:19:25 »
Might be worth posting to Dropbox or similar a single flat as they show defects pretty clearly. I would suspect you have two column defects, they don't always show up as fully black.

Chris

Offline astrodoc71

  • PixInsight Enthusiast
  • **
  • Posts: 92
    • View Profile
Re: Difficulty removing bad column
« Reply #10 on: 2017 April 10 05:53:35 »
Didn't really see it in the flat but here it is if you want to look at it:
     https://www.adrive.com/public/2m9Yu8/HASH2%20249_308PAWest.00052.zip
Thanks
Dave

Offline chris.bailey

  • PixInsight Addict
  • ***
  • Posts: 235
    • View Profile
Re: Difficulty removing bad column
« Reply #11 on: 2017 April 10 11:16:51 »
Very Strange!

From that flat I would say columns 1377 and 2105 are the suspect ones and are removed from running CC.

Chris