Debayering One-Shot Color data

[dpaul]

Can someone assist -

I've just started using the ATIK Horizon color camera (already have the mono and having good results with LRGB Imaging).
Im having problems converting the one-shot color data from raw to color via debayering.

The sensor is a Panasonic MN34230 4/3" CMOS.

This is the order of play which I 'think' is the correct procedure:  I calibrate the raw light frames then debayer each frame. Then I do star alignment and integration of the images. When debayering it doesn't recognise the CFA  (Error: Unable to acquire CFA pattern information: Unavailable or invalid image properties). However I switch from auto to a specific pattern (bayer/mosaic pattern) it then works. I don't currently know the pattern for sure but I think its BGGR (checking that directly with Atik). My understanding is that Auto used to work on earlier Pixinsight version but now it doesn't (not sure if that's correct?) - either way that's an issue.

So after registering and integrating the frames, the resultant RGB combined image seems to have little or no color. It doesn't matter whether I stretch the data, play around with the curves, try color colibration, etc, its just generally gray. The only coloor that seems to appear is the background gets light color when saturation is pushed to the limit. I've also tried looking at an individual debayered frame before intregration and that's the same. I've also tried non-calibrated frames and still the same.

I also tried every pattern in addition to BGGR and still no better (some gave a green hue overall) - still no general colors.  I was using VNG as the demosaicing method but bilinear and superpixel didn't work either.

Attached are the statistics (clippe dand unclipped) for a single raw frame that was debayered with BGGR / VNG (no calibration or registering) - doe sthis data look ok?

I have a 30'' scope and the frame length was 15 seconds - this is a lot of data with this aperature and proves more than enough with LRGB imaging so I'd be suprised if its a weak signal. Maybe I'm doing someing fundamentally wrong when saving the data but I'm saving RAW / FITS so I'd be surprised.

Thanks in advance for any assistance.

David

[dpaul]

Just one small correction - I meant to say AUTO or not AUTO working ''isn't an issue''

[bulrichl]

Hi David,

the general approach that you described is correct.

Can you supply a calibrated light frame? I'll take a look at it.

Maybe the hints that I gave in the following thread can help you to find out the right CFA pattern:
https://pixinsight.com/forum/index.php?topic=11571.0

Bernd

[dpaul]

Hi Bernd,

Thanks for the reply -

I can send a calibtared lightframe but how do I do this on the forum, the file size is 16Mb?

David

[bulrichl]

An upload of big files to the forum is not possible. You would have to establish an account at a filehoster like dropbox or something similar and supply here a link to that specific file.

Bernd

[dpaul]

Hi Bernd,

I'll leave it for now - I'll see what comes back from Atik, that may help me.

Thanks

David

[Warhen]

Auto won't work for OSC FITS, which doesn't report its Bayer pattern. It's for DSLR RAW files, which generally do provide the pattern. David, you certainly do NOT need to debayer individual files! This can be done in batch by the new Debayer process, or in the background when using the BPP script. Regardless, you MUST establish the correct Bayer pattern beforehand.

[dpaul]

Hi Warren,

Thanks for this, I will only be debayering the integrated FITS frames.

The main problem was trying to debayer 2x2 binned files, I need to work on 1x1 only.
Not tried it yet, been having fun with LRGB imaging with a mono camera and its been snowing for a week in the UK.

Will try again with OSC soon

Thanks

David