Problems with OSC calibration/integration

[JGMoreau]

Hi, After many hours remaking masters calibration frames from raw,
making triple sure every parameters in preprocessing script is correct,
so with integration module, i still get the same result :

http://www3.cablevision.qc.ca/astro/Calibration_Integration_Problem.jpg

A comparaison with DSS included 

If you want to try for yourself, some images/masters are in Moreau folder on the file sharing site.

Please someone help me sort this out

Jean Guy

[pfile]

what is the correct bayer matrix for this camera? also your .fit files look to already be calibrated, is that correct?

[Cyclop_si]

Maybe you just have to do Cosmetic Correction at the begining, before Celibration?

[JGMoreau]

Thanks for the help.

My camera is ST2000 XCM with bayer matrix BGGR.

The M61 frames are .fit straight from the camera, without any calibration.

However, this morning, after the latest core update, a quick test with the
debayer module looks very promising

Some more fun today it seems, what a great adventure and fantastic peoples
on this PixInsight cruise of the cosmos.

A very special merci to our captain Juan.

Jean Guy

[JGMoreau]

Well, early morning enthusiasm turned into disappointment as the
persistant bayer matrix is still present in my fimal integrated image

http://www3.cablevision.qc.ca/astro/Bayer_horror.jpg

Some more patience is call for i guess...

Back to DSS.

Jean Guy

[pfile]

i dont understand this, i debayered your images successfully. i happened to use superpixel debayering but any algorithm should work.

i wonder if something is getting confused re CFA vs. mono.

[Juan Conejero]

Hi Jean,

Your images can be demosaiced without problems with the Debayer tool, using the RGGB mosaic pattern and any debayering method. I guess you have simply a FITS orientation problem. Please do the following:

- Open the Format Explorer window.

- Double click the FITS format item.

- On the FITS Format Preferences dialog, select the upper left corner (up-bottom) as the coordinate origin (Miscellaneous Options section).

- Click OK

Now your images will be debayered without problems.

Note that in the latest versions of the Batch Preprocessing script (0.5) I have included an 'up-bottom FITS' option that can be enabled (it is enabled by default) irrespective of the FITS Format Preferences setting.

Hope this helps.

[JGMoreau]

Hello Juan,

Thanks for replying promptly.

I agree that one debayer image seems ok, but the result of integrating
50 of them bring back a remnant of the bayer pattern :

http://www3.cablevision.qc.ca/astro/Debayer_problem.jpg

Oh No! not the Moreau effect!   

DSS uses the ADH interpolation process, and i dont get this effect,
so maybe its a clue?

Jean Guy

[Juan Conejero]

Hi Jean,

I haven't seen anything similar before. This result is extremely strange after integrating 12+ frames. Are you sure that your master calibration frames are correct? Have you rebuilt them from scratch with the correct FITS orientation?

[JGMoreau]

Hi Juan,

I finaly nailed it down, with your help of course

Thank you very much.

Here is what worked for me, and my ST2000XCM camera:

All calibration masters made using 32 bits FITS format right from camera, and

Upper left corner option for FIT format preference.

Batch preprocessing works like a charm.
Used RGGB pattern as you said.
Image Integration as per settings in this image :

http://www3.cablevision.qc.ca/astro/Youhoo.jpg

Tremendous gain in image quality!!!

Happy Jean Guy

[JGMoreau]

One more comment i would like to make.

The M61 images in Moreau_files are .fit and the master calibration frames are .tiff

Was this the source of my calibration problems?

If so better be carefull when saving your master frames

Also, i tried to delete the Moreau folder and files, but do not have permission...

Jean Guy

[Nocturnal]

Not likely. If PI saved both TIFF and FIT masters it would have re-loaded them as a normalized 0,1 range. Still I recommend using a single format. I use TIFF although the FIT keywords can be useful.

[georg.viehoever]

Was this the source of my calibration problems?

Maybe loading via tiff creates a differently rotated image. I saw it a while ago that using fits I got mirrored results. I am standardizing on fits now.
Georg