Author Topic: Zwo 071 debayer  (Read 1186 times)

Offline Ken82

  • Newcomer
  • Posts: 16
Zwo 071 debayer
« on: 2019 September 15 02:17:52 »
Ive been using a dslr until now with good results. Decided to change to a cooled 071 and i appear to be doing something wrong as my images are massively green. I believe RGGB is the correct debayer setting ? Im assuming the calibration procedure is the same as it is a dslr sensor in effect.

The other slight concern is that my master dark doesnt appear correct ?


 Not calibrated   https://drive.google.com/file/d/14b8a_Fub46Sc8nu4MqaETQRjFdTKlOH-/view?usp=sharing

 Calibrated, debayered      https://drive.google.com/file/d/1eZATcCj6ND0gm1NiDV0xvO8VHJsLfnRj/view?usp=sharing

master dark    https://drive.google.com/file/d/1fu9aTnf5NL93h7EExG-p7jfEvVWNM8cV/view?usp=sharing


Thanks for any help,
ken

Offline pfile

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 4729
Re: Zwo 071 debayer
« Reply #1 on: 2019 September 15 08:05:02 »
you are OK - this is normal. it happens because there are 2 green pixels in the bayer matrix and hence the SNR of the green channel is 1.4x that of the other channels. this in turn "fools" STF when it is computing a stretch with the channels locked.

if you want to preview what the file should look like, just unlink the channels in STF by clicking the chain icon at the top left of the STF interface and recomputing the stretch (nuclear button.)

the proper solution to this is to run BackgroundNeutralization with the threshold set high enough to capture the background in the green channel, or run PhotometricColorCalibration making sure to set the threshold similarly so it can neutralize the background.

rob

Offline Ken82

  • Newcomer
  • Posts: 16
Re: Zwo 071 debayer
« Reply #2 on: 2019 September 15 09:43:21 »
Thanks Rob!

I will try out that PhotometricColorCalibration later this evening. Odd that it doesnt do this with a canon 6d which is effectively the same RGGB.  Does the master dark look correct to you as it does appear to have a large dark cross to me ???

thanks
ken

Offline pfile

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 4729
Re: Zwo 071 debayer
« Reply #3 on: 2019 September 15 09:47:02 »
i didnt look at it - i will check.

it is weird that you never saw it with the DSLR because the green cast is certainly prevalent there. were you using any kind of an LP filter with the DSLR?

rob

Offline Tertsi

  • Newcomer
  • Posts: 16
Re: Zwo 071 debayer
« Reply #4 on: 2019 September 15 12:13:00 »
Hi Ken,

I have same camera (ASI071MC-pro cool) and I have a problem in calibrating files. Post is here: https://pixinsight.com/forum/index.php?topic=12850.msg84011#msg84011. May I ask details about your flat frames? What kind of light source do you use for flats and what are exposure times and camera gain and offset settings?

Sincerely,
Tertsi


Offline Ken82

  • Newcomer
  • Posts: 16
Re: Zwo 071 debayer
« Reply #5 on: 2019 September 15 14:11:00 »
Hi Tertsi,

I havent taken any flat frames with this camera yet as this was just a test but your thread is very interesting. I was under the impression the flat frames were just used in greyscale so ive never debayered one to be honest.

Maybe thats my next hurdle in the transition to the zwo 071. Ill be sure to read your thread some more!

Cheers ken

Offline Tertsi

  • Newcomer
  • Posts: 16
Re: Zwo 071 debayer
« Reply #6 on: 2019 September 15 21:55:21 »
Hi Ken,

Generally speaking, flat frames are needed in light frame calibration.

After little bit of more experimenting, I think there is a problem in PI calibration, because using same set of files (light, bias, flats) in Deepskystacker I get expected final image with nice histogram.

Sincrely,
Tertsi

Offline pfile

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 4729
Re: Zwo 071 debayer
« Reply #7 on: 2019 September 15 23:18:18 »
there's no problem in PI calibration save for the fact that there is a single scaling factor applied to OSC flats. and it isn't really a problem, because the process of flattening an image is simply dividing the flat into the light, which is a linear operation. you can then multiply any channel of your image by any amount and the result is still linear, and if the amounts are chosen carefully, you will undo any cast created by the flat. thus if your image is left with a color cast due to flattening, it is truly a simple matter of multiplying two channels by some factors to restore the balance. nothing is lost in this process. in fact, this is exactly what the various color calibration processes in PI do - compute scaling factors and multiply each channel by those factors.

if you endeavor to make your flats grey, then the process of flattening the image in PI will not change the color balance. however, it's just a fact that all astronomical images need some kind of background neutralization and color calibration regardless of what the flat looked like. the atmosphere, the different transmission percentages of your filters and the quantum efficiency of your sensor at particular filtered wavelengths all conspire to mess up the color balance of your image anyway. so no matter what you do, you can't rely on the colors coming straight out of the camera. you need to perform color calibration.

to ken's question - yes, flats, darks and bias are all applied to the light while the image is still in CFA mode. debayering a flat is really only useful to check what the flat histogram looks like - and generally what you are looking for is that all 3 channels are pretty much exposed to 1/2 the well depth of the sensor at the center of the image. if there is a color imbalance, it's not fatal as described above, but a serious color imbalance can be a problem in the sense that the weakest channel might not have enough signal, and therefore the SNR of the stacked master flat in that channel is poor. this adds more noise to your calibrated frame than necessary.

rob


rob

Offline Tertsi

  • Newcomer
  • Posts: 16
Re: Zwo 071 debayer
« Reply #8 on: 2019 September 15 23:59:12 »
there's no problem in PI calibration save for the fact that there is a single scaling factor applied to OSC flats. and it isn't really a problem,

But PI calibration totally messes up my images as one can see from the histograms. I have never seen this before. Is that possible you to double check image calibration and stacking process? I have used BatchPreProcessing script. I can make my light, flats and bias frames (10 of each) available for download. I would really appreciate help on this.

Edit: This problem solved with setting "UP-Bottom-FITS" ON in BatchPreProcessing script.
« Last Edit: 2019 September 16 07:42:44 by Tertsi »