Pixelmath R, G and B to obtain L

[Falco251]

Hi!

I was wandering if it makes any sense, after capturing a set of R, G and B frames of a target, to do this:
- calibrate and register all frames as usual;
- perform a sum in pixelmath of R1+G1+B1 = L1, R2+G2+B2 = L2, ...
- integrate L1, L2 etc. together with actual luminance frames captured with the Lum filter.

I understand that the SNR of a pure L frame is higher, but could the use of these synthetic L frames with the actual L ones provide an increase in the amount of signal in the final master L?

In simple words: say I capture 12 frames with each filter (12xL, 12xR, 12xG, 12xB), same binning and duration for each image. Will the luminance be better if:
1) I integrate just the 12 L frames;
2) I integrate 12 L frames + 12 "pixelmathed" (R+G+B) frames (total of 24 frames)?

Thank you

[jkmorse]

Depends on what your binning was for the RGBs.  If they were shot with the same binning as the Lum, then you can combine and get a little kick.  If they are 2x2 and your lums are 1x1, then you will actually hurt the resolution by combining. 

[Falco251]

Thank you for your reply!
Sorry I forgot to mention that, I'm talking about images with the same binning.

At the moment I use a short focal length refractor that provides a wide FOV.
I captured a set of 15 L frames of M31 in bin 1 and a set with R, G and B in binning 2, but I think that the sampling of the color images is too low and I'm planning to take more R, G and B picture in binning 1.
I'm fairly happy with the L image, but I was wandering if adding the synthetic L generated with PixelMath might provide a little more SNR to the Lum as well.

[Herbert_W]

Hi,
 
there is the possibility to create an optimal luminance image from R/G/B frames with the module ImageIntegration - also for OSC/DSLR images.

Statement from Juan:


... it is very easy to implement with the ImageIntegration tool. If you work with an OSC camera you have to split your color image first with the ChannelExtraction tool, and save the individual R, G and B images as FITS files. Open ImageIntegration and select the four files. Then leave all tool parameters by default (you can click the Reset button to make sure) and click the Apply Global button. The relevant parameters are as follows:

- Combination = Average
- Normalization = additive with scaling
- Weights = Noise evaluation
- Scale estimator = iterative k-sigma
- Generate integrated image = enabled
- Evaluate noise = enabled
- Pixel rejection = No rejection
- Clip low range = disabled
- Clip high range = disabled

You can make several tests with different scale estimators and select the one that yields the highest noise reduction. The integration result is the optimal luminance image that you can treat in the usual way (deconvolve it if appropriate, stretch it to match the implicit RGB luminance, combine with LRGBCombination, etc).

Note that the same procedure can be used to compute an optimal luminance for an RGB image, that is without an additional L image. In this case just use one of the RGB channels as the reference for integration.

Source: http://pixinsight.com/forum/index.php?topic=6042.15  page 2 #17

Best regards!

Herbert, Austria

[Falco251]

Thanks for the reply! Last night I captured a few hours of RGB in bin 1 of M31 and I'm looking forward to try this out.

I have a new question though. What is the best procedure:
1) pixelmath R1 + G1 + B1 = L1, R2 + G2 + B2 = L2, ... and in the end integrate the resulting L1, L2, etc.
2) integrate all R, G and B and then pixelmath the result to obtain the synthetic L.

I guess that if I have other real L images the way to use the RGBs is option 1), but if only RGB data is available then 2) is better. Am I right?

[oldwexi]

1) Usually i integrate all calibrateted R,G,B with process ImageIntegration in ONE step to a synth L
     this should give better SNR
2) i am a PixelMath fan, but in this case is see no need for it.

Gerald

[Falco251]

I applied the procedure indicated and it seems to work quite well. Can I perform DrizzleIntegration in the same way as well?