Like most websites, SGL uses cookies in order to deliver a secure, personalised service, to provide social media functions and to analyse our traffic. Continued use of SGL indicates your acceptance of our cookie policy.

# RGB + HA

## Recommended Posts

Normally using pixel math in pixinsight i dont have any issues combining rgb with ha but on this occasion its not working as well as i would like. Ive followed the procedure as normal from the light vortex tutorial but still not great. Any guidance would be really appreciated.

Thanks Ken

Both images are stacked and have had background extracted but no processing.

HA

RGB

##### Share on other sites
On 17/10/2019 at 23:21, Ken82 said:

Normally using pixel math in pixinsight i dont have any issues combining rgb with ha but on this occasion its not working as well as i would like. Ive followed the procedure as normal from the light vortex tutorial but still not great. Any guidance would be really appreciated.

Thanks Ken

Both images are stacked and have had background extracted but no processing.

HA

RGB

I tend to follow the process steps as advised by Adam Block's PI tutorials  - see https://adamblockstudios.com/categories/PixInsight if you want to watch a video of all this.

1. Align Ha image with the Red channel of the RGB image

Given that your Ha data is actually encoded as RGB, you first need to extract the Lum. To do this correctly, apply RGBworking space, with all the parameters set to 1 to the Ha image and the RGB image. Then extract the CIE L* from the Ha (RGB) image and use channel extraction to extract the Red channel from the RGB image.   Now apply star alignment to the Ha (Lum) and the Red channel, taking the Red channel as the reference.  You now have unstreched but aligned Ha and RGB images.

2. Make stretched versions of the Ha (lum) and RGB image

Use the STF and HT functions until you are happy with images - I just used the auto stretch function.

3  Blend the Ha and RGB red channel together

Before you do this you need to manipulate the Ha data by:

a) To make this process a little easier, perform a linear fit on the Ha information using the red channel as the reference. This makes the intensities of the two images similar, making then easier to blend together.

b) Given that you don't wish to put the Ha into the stars, one option is to use PI's starnet++ to make a starless image of the Ha which you can do by simply applying starnet to the stretched Ha image.

c) Run TGV denoise on the Ha starless image - this is to minimize the chance of raising the noise floor of the red channel when you do the subsequent blending.

d) You now need to create an image that only has the brightest parts of the Ha information.  This involves black clipping the Ha data so that you don't end up creating a red cast on the RGB data. This is the critical step in the process.

e) You now have two images: a starless Ha black clipped image which only has the brightest parts of the Ha image and the Red channel of the RGB image.

f) Use the script Blend script (download at https://www.skypixels.at/pixinsight_scripts.html) to blend the two images together you end up with an Ha enhanced red channel. Select the screen blending mode on the script.

4. Recreate the RGB by taking the extracted Blue, Green and Ha enhanced red channel.

You are done.

From your data, I ended up with the following blended image + after application of SCNR green. I didn't do any subsequent processing eg noise reduction/colour enhancement since I thought you would want to see if just with the Ha blend (which looks quite good to me).

A few minor points that are nothing to do with Ha blended that you might find useful - these only show up if you zoom in.

1. If you look at the Fits header information of the source images, something appears to be incorrect with your image integration. On the RGB image you appear to have zero rejected pixels from the integration of 10 images.  you can see the effect of this if you pixel peek eg a satellite trial is visible.  You should be ending up with 1 to 2 %. On the Ha image you have more than 10% rejection, from a stack of 42 which seems a bit aggressive to me.

2. Overall your stars look good but on the right hand side that are elongated, I presume this is camera tilt.

Alan

##### Share on other sites

Thanks Alan really appreciate your detailed response !

I would normally extract the CIE L (red channel)  but you have gone into further detail after this which is different to my current process.

thanks ever so much I will look into implementing what you have suggested.

ken

## Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

×   Pasted as rich text.   Paste as plain text instead

Only 75 emoji are allowed.

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×