debayer

[brrttpaul]

Using a OSC at the moment but hoping to upgrade to mono very soon so I bought pixinsight, so ATM im using Maxim for my capture and my pre processing but I was thinking of using pI for my preprocessing and stacking, now using a mono all well and good as theres lots of video tutorials out there but what about when using a OSC with regards to flats? in maxim i have to click the apply boxcar filter to debayer the flats but what would I need to do in PI ( of course once I go mono I wont need to do it but I thought I might get some practise in if we ever have a clear night in the meantime.

[Star101]

I find it best to debayer your lights and flats before any other processing.

Debayer by SCRIPT/BATCH PROCESSING/BATCH DEBAYER. Just add all your lights, select the output and the bayer matirx and run. Then do same for flats etc. One could do all the files at once but I prefer keeping the files separate. 

There are a few ways to stack images in PI. I prefer the one by Richard Bloch Calibration and Stacking in PixInsight 1.8. Although this one is done for DSLR images, OSC CCD is close. The only caveat I know is in the video at 16:06 UNCHECK the OPTIMIZE box. I had problems with colour being totally wrong when this  box is checked. It appears attempts to debayer no matter if you have or not. And it appears to guess the matrix used. Best leave unchecked    ( unless someone else can suggest different! )

https://www.youtube.com/watch?v=zU5jJgjKuQQ&t=452s

Then process the final image using another video by the same.

https://www.youtube.com/watch?v=6HUOe4R5Hng&t=162s

[ecuador]

Theoretically calibration would benefit from performing it before debayering - although the benefit might be small. I haven't used PI for calibration, doesn't it have such an option?

[brrttpaul]

Thx for the info guys, guess it was staring at me in scripts/batch processing/ batch debayer.  the Richard Bloch pix 1.8 is the video i have been looking at funny enough and trying to go down that route, thx for the help

[brrttpaul]

At thepoint of asking the "this weeks idiot question award". But why arnt darks or bias debayered? surely the chequered effect will be on there also, the reason im asking is they will load on the debayering programme on PI

[AKB]

Darks and Bias know nothing about the Bayer matrix (since it is dark!) and their effects are on a per-pixel basis.  A de-Bayered image of a bad pixel (or high bias) will affect at least three other pixels as well - this is not what you want from calibration frames.  So these corrections should be applied before the rest of the processing chain.

Personally, I prefer to use a bad pixel map, rather than darks.  Then you're not in danger of removing the bias twice.

 

[brrttpaul]

Thanks for the reply, I was just wondering why thats all because the chequer effect in lights and flats are smoothed out thought maybe the same for darks/bias thx

[richyrich_one]

Calibration frames are applied to light frames undebayered and then only the calibrated lights are debayered before stacking.

https://pixinsight.com/forum/index.php?topic=7627.0

[Waldemar]

Maybe this will help? : http://www.harrysastroshed.com/pixinsight/pixinsight video html/Pixinsighthome.html

[wimvb]

+1 for @richyrich_one's reply.

Image calibration should always be done with bayered (unprocessed) images, never debayered images. If you use PixInsight batch process to calibrate and stack, you load all raw image files as you would in deep sky stacker. Then check the box marked CFA image files, and choose the correct bayer matrix (RGGB, BGGR, RGBG, etc). If you don't know your bayer matrix and can't find a reference on the internet, try this:

Take an image in daylight and apply the debayer script in PixInsight. Choose various alternatives for the bayer matrix, until the colours look right.

Hope this helps

[iapa]

On 27/02/2017 at 19:08, brrttpaul said:

Using a OSC at the moment but hoping to upgrade to mono very soon so I bought pixinsight, so ATM im using Maxim for my capture and my pre processing but I was thinking of using pI for my preprocessing and stacking, now using a mono all well and good as theres lots of video tutorials out there but what about when using a OSC with regards to flats? in maxim i have to click the apply boxcar filter to debayer the flats but what would I need to do in PI ( of course once I go mono I wont need to do it but I thought I might get some practise in if we ever have a clear night in the meantime.

 

On 27/02/2017 at 19:30, Star101 said:

I find it best to debayer your lights and flats before any other processing.

Debayer by SCRIPT/BATCH PROCESSING/BATCH DEBAYER. Just add all your lights, select the output and the bayer matirx and run. Then do same for flats etc. One could do all the files at once but I prefer keeping the files separate. 

There are a few ways to stack images in PI. I prefer the one by Richard Bloch Calibration and Stacking in PixInsight 1.8. Although this one is done for DSLR images, OSC CCD is close. The only caveat I know is in the video at 16:06 UNCHECK the OPTIMIZE box. I had problems with colour being totally wrong when this  box is checked. It appears attempts to debayer no matter if you have or not. And it appears to guess the matrix used. Best leave unchecked    ( unless someone else can suggest different! )

https://www.youtube.com/watch?v=zU5jJgjKuQQ&t=452s

Then process the final image using another video by the same.

https://www.youtube.com/watch?v=6HUOe4R5Hng&t=162s

From this, I take it that the 'boxcar filter' in Maxim is a mechanism to de-Bayer a colour image?

I have seen the term is several places, but usually in a workflow, but not a description of what it is actually doing

[Star101]

On 06/03/2017 at 22:57, wimvb said:

+1 for @richyrich_one's reply.

Image calibration should always be done with bayered (unprocessed) images, never debayered images. If you use PixInsight batch process to calibrate and stack, you load all raw image files as you would in deep sky stacker. Then check the box marked CFA image files, and choose the correct bayer matrix (RGGB, BGGR, RGBG, etc). If you don't know your bayer matrix and can't find a reference on the internet, try this:

Take an image in daylight and apply the debayer script in PixInsight. Choose various alternatives for the bayer matrix, until the colours look right.

Hope this helps

I have never seen the (RGGB,BGGR,RGBG etc) option when I select CFA in PI. If you could tell me where to find it I would be very appreciative.I have tried ALL three options and they all have the wrong colour settings for my ATIK 4120EX camera. So, I debayer my lights and flats then run the calibration etc with OPTIMIZE UNCHECKED. Works great this way for me

 

[wimvb]

We're talking about 2 different processes in PI. I was referring to the BatchPreprocess script, which has a tic box called CFA images (Global options), and a deBayer process. In the deBayer process, you set the Bayer matrix I wrote about in my previous reply.

The 'manual' process (see this link: http://www.lightvortexastronomy.com/tutorial-pre-processing-calibrating-and-stacking-images-in-pixinsight.html#Section1 for a workflow) doesn't DO deBayering during the calibration process, it only DETECTS if an image is monochrome or has a Colour Filter Array.

While it is very much possible to deBayer the master flat and light frames before calibration, this isn't the intended workflow. For one thing, if you have hot pixels left in the masterflat, they will become smeared out during the deBayer process, and affect several surrounding pixels in the calibrated light frame.

Here's more information on dark optimization: http://pixinsight.com/forum/index.php?topic=4161.msg29059#msg29059

And here's an article about deBayering: http://www.stark-labs.com/craig/resources/Articles-&-Reviews/Debayering_API.pdf

[Star101]

I see, yes, you are processing one of the other ways in PI. 

I will try this method. Maybe it will work better for me. 

I am still learning PI, Its a excellent program but has so many ways to do things. If I can improve my images more, I am happy to try

Thanks for clarifying