Author Topic: newbie issue with dark frame substraction  (Read 2977 times)

Offline rpineau

  • Newcomer
  • Posts: 5
newbie issue with dark frame substraction
« on: 2013 November 24 17:39:47 »
Last night I was testing a new telescope (AT 12 RC) and tried to image M74.
I did about 10 frame of 600 seconds each and 3 dark frame on 600 seconds each.
In PixInsight, I took the 3 dark frame and integrated them (Sigma Integration), saved them as one master dark frame.

I then proceed to calibrating the light frame (using ImageCalibration). Selected all my light frames , checked only Master Dark as I didn't do any flat or bias (this was still just for testing), added all my light frame and applied. Only Optimized was check for the dark frame.

I then did the registration with StarAlignement, select the first light frame as the reference and added all the other (add files).
Once aligned I went back to integration, selected all the aligne frame (adding in the reference one).

I enable the ScreenTransferFunction and noticed some artifact on the images : dark pixels that look like the one from the dark frame with the image drift applied !!
you can see the (quickly) processed fits file here : http://www.rti-zone.org/files/M74_test.fits
and a PNG version for a quick look : http://www.rti-zone.org/files/M74_test.png

So.. what am I doing wrong here ?
I was assuming that when PixInsight was removing a hot pixel from the frame it was replacing it with the average (or other interpolation) of the nearby pixels. If that's not the case how do I do this ?

Thank you
Rodolphe


Offline TobiasLindemann

  • PixInsight Addict
  • ***
  • Posts: 126
    • Trackingstation
Re: newbie issue with dark frame substraction
« Reply #1 on: 2013 November 24 20:14:28 »
Hi Rodolphe,
it seems you used a ccd-camera. Why do you use only 3 darks for the masterdark? I would do 10 darks at least. If you have a controlled cooling you can do your darks when ever you want, right?
Interative, or Winsorized Sigma clipping works better with at least 10-15 Images. For only 3 images "Percentile clipping", or "Averaged sigma clipping" works better. It is hard what to reject if you have only 3 images, you know.
If you have still black pixels in your calibrated images you can use "cosmetic correction" to get rid of them.
I have tried it on your integrated image and it worked well http://www.tracking-station.de/deepsky/M74_test_cc.fit, but I think it will work even better with the single linear images. I think it can also reject more (or all) of the hot pixels in the single linear images.

Greetings
Tobias

Offline rpineau

  • Newcomer
  • Posts: 5
Re: newbie issue with dark frame substraction
« Reply #2 on: 2013 November 24 21:50:28 »
Thank you for this answer and the test you did.
Yes the camera is a cooled CCD (SBIG STF-8300M). So I'll take more dark to at least match the number of frame I have (only 6 in this case).
I'll try to redo the processing by following your advice (I'm just starting with PixInsight so it might take me a few iteration to get the process down.).
I'll try the "cosmetic correction" on the individual frame before integration.

Thanks again for your help on this.

Offline Geoff

  • PixInsight Padawan
  • ****
  • Posts: 908
Re: newbie issue with dark frame substraction
« Reply #3 on: 2013 November 24 21:55:51 »

I'll try the "cosmetic correction" on the individual frame before integration.

Harry shows how to do this in the following video http://www.harrysastroshed.com/preprocessing.html
Geoff
Don't panic! (Douglas Adams)
Astrobin page at http://www.astrobin.com/users/Geoff/
Webpage (under construction) http://geoffsastro.smugmug.com/

Offline rpineau

  • Newcomer
  • Posts: 5
Re: newbie issue with dark frame substraction
« Reply #4 on: 2013 November 28 12:17:35 »
Thank you all.
I followed the video tutorial and was able to get much better result.
Thanks again for all your help.
Regards, Rodolphe