Ha debayer or "super pixel " ?

Hi Andy @stargazer,

Please have a look at the information in my presentations on this subject:

https://www.astropixelprocessor.com/presentations/

https://s3.eu-central-1.amazonaws.com/apastropixelprocessordl/APPOverviewOfCurrentAndFutureFunctionality-final.pdf

Pages 27,28,29

APP has a special h-alpha debayer algorithm.

If you select it for debayering then APP will immediately show you the monochrome H-alpha data withous loss of resolution.

No need to do superpixel and/or splitting of the channels with APP.

It works totally automatic and thus the workflow is very simple, only set the debayer algorithm in 0) Raw/FITS to the H-alpha debayer mode 😉

Images show a crop of a Canon 6D CR2 frame shot using a H-alpha filter. Second images is after calibration. Third is calibrated and with the H-alpha debayer algorithm applied.

APP is the first application that has a simple and direct solution to your question and the H-alpha debayer algorithm is the best possible workflow:

• No loss of resolution
• Therefore more precise registration of your data ( with resolution halved in the x- and y-axes, star location calculation is much worse !)
• No need to split the channels ( APP will immediately show the monochrome H-apha data in the red channel)
• Since there is no loss of resolution, drizzle is not needed to restore the resolution and therefore your end data has lower noise. (Drizzle is a noise injector, so if you can avoid using it, it’s always better)

Let me know if you have more questions.

Kind regards,

Mabula

It is really good. I select it before I do anything when dealing with hydrogen Alpha is that correct? After selecting that I make my mastet flats, darks, and bad Pixel Map[once).  And here's where I may be making a mistake because I also load the master bias with the lights. All this is done while hydrogen Alpha debayer selected.

@gregwrca,

Using this in combination with data calibration is not different than normal RGB processing with a normal debayer algortihm. Bayer CFA data needs to be calibrated on the CFA pixels before debayering, so it works exactly the same as normal RGB data.

Debayering, and thus H-alpha debayering, is done automatically after data calibration 😉

Mabula

Thanks for the reply.

What kind of  interpolation is there between pixels , to "fill in "  the green and blue pixels , and adjacent red pixels ?

Hi Andy @stargazer,

Well, debayering of H-alpha data can really only be done using the CFA pixels in the red channel. The information in the green and blue CFA pixels is almost zero.. althought the H-alpha signal slightly leaks through the Bayer CFA pixels.

So if you choose the H-alpha debayer algortihm, the Red CFA pixels are used to interpolate the other red pixels.

Nothing is done in the green and blue channels and only the red channel is shown (therefore it's monochrome), since that captured the H-alpha signal.

The actual interpolation is similar to regular bilinear interpolation. More advanced interpolation isn't really possible since those kind of interpolations are usually based on clear signal in the green channel and that helps in constructing the red and blue channels with more accuracy.

If you use the OIII debayer alogrithm with an OIII filter in fronft of your DSLR, than a more advanced interpolation is used, since the green channel shows siginificant information, as does the blue channel. The Green and Blue channels are debayered with more advanced interpolation and both channels are then combined to show the monochrome OIII signal.

Kind regards,

Mabula

Hi Mabula,

Thank you for clarifying  matters.

Best Wishes

@stargazer 😉