PixInsight Forum (historical)
PixInsight => Tutorials and Processing Examples => Topic started by: vicent_peris on 2007 December 22 15:07:57
-
Hi,
this is a great forum! I present today the end result of our discussion about H-alpha / RGB combining methods. You can now learn how works my new algorithm to integrate narrwoband with broadband images. Juan has formatted the first part and it is accesible from the PixInsight's webpage. You can also follow this direct link:
http://pixinsight.com/tutorials/STD/narrowband/theory/en.html
As the introduction of the article says, this new method has several advantages:
* Fully controllable enhancement of line emission images.
* Accurate color representation of both line emission and continuum emission objects.
* Minimum signal-to-noise degradation in narrowband data.
* Effective star brightness reduction.
* Fully automatable workflow with consistent results.
Remember that we will implement this algorithm as a PixInsight module ASAP.
I must give many thanks to Jack Harvey for giving me the permission to work with his wonderfull data. This algorithm has been devised thanks to it collaboration.
The ongoing work... For the moment, this firs part is traduced to Spanish and will be uploaded ASAP. I'm going to write now the second part of the article: the practical one.
So... I must buy now a CCD camera! :lol: :lol: :lol:
Well, hope this will be usefull for the astrophotographic community.
Best regards,
Vicent.
-
Hi Vicent,
Great tutorial and very useful for us. We are currently working with narrowband data, and now our next planned target is the horse head region (or is it the cone?, one or both :lol: ) and we are planning to do it with Ha and RGB filters, so we'll have the opportunity of try your tutorial :wink: . Thank you!
Regards,
Oriol and Ivette
-
Hi Vicent,
Thank you very much, it is a very good tutorial with bright ideas. I have to try it with some of my images.
By the way, I sometimes get OIII images as well as Halpha and RGB and I use to mix Halpha with red and OIII with both green and blue. I think that I can apply the first part of your tutorial to mix the OIII, but do you think I can use later the ideas of the second part of the tutorial to mix the H-alpha as the luminance?
Best regards,
Antonio.
-
Hi Antonio,
thanks! I've sent you a PM.
Vicent.
-
Fantastic job, I'm really impressed.
-
Vicent,
Another fascinating tutorial, and it was great to track your thinking along the way. But, out of curiosity (that's a kind way to say my own "ignorance" :D ), I have a question: could you also subtract the Ha from the Red to get the same map? Not sure why I'd want to, but would I still be left with just the broadband components of the Red channel to denoise?
Again, lots of of new stuff to work into images,
Larry
-
Hi Larry,
it works too, but the division has a great advantage, because it recovers a lot better the fainter stars inside the nebulas.
Regards,
Vicent.
-
Very interesting work Vicent, thanks for sharing :wink:
Regards
Jordi
-
Hi,
I have great news.... seems that the noise reduction step can be enormously simplified. Because I'm experimenting with the K-sigma noise thresholding function of the A Trous Wavelets, and this seem to work well!
It would be great to incorporate a star protection option to the noise reduction (the same as in ACDNR tool). Would it be possible, Juan??
Vicent.
-
Hi Vicent,
Really great news! I think we could add an edge protection feature (similar to ACDNR) to a wavelet-based k-sigma thresholding routine. That's an excellent idea! I'll have to experiment with it...
-
Seem that it's the way... but I'm seeing that it transfer some large scale noise to the narrowband image... we must think about this noise reduction, surely with some modifications it will work.
Vicent.
-
Good work Vicent. Thank you very much.
-
Very interesting idea!
Could you talk a little about how you normalized or offset the Red and H-alpha images before doing the division operation to make the continuum map? Any other required normalizations or matching background offsets required for the technique?
-
¿Se puede trabajar con las imágenes antes de hacer transformaciones no lineales, es decir, según salen tras el apilado?
Yo es que no consigo nada, la imagen se va degradando con las iteraciones y no logro aflorar las estrellas. Podéis explicar el proceso de iteraciones más detenidamente. Creo que me estoy haciendo un lío.
Un saludo,
Sergio
-
¿Os habeis dado cuenta de que habeis 7 españoles hablando en inglés?, estaria bien el tutorial en castellano, el traductor automatico es una birria :cry:
-
caliu, si quieres lo vamos comentando entre nosotros en español y nos vamos aclarando entre los dos. :D
Sergio
-
Vicent,
I was attempting your procedure last night and for some odd reason, when I did the division (i.e. Red/Ha), I only got about six of the brightest stars in the continuum map. The red image has probably got over a 100,000 stars while the H-alpha image as around 20,000-30,000 stars. Although both images could be stretched more, I backed off on the stretching for the preservation of star colors. In addition, the continuum map looked very dark, I really had to stretch it before it looked similar to your background. Working with very wide-field images is proving difficult. Any suggestions would be greatly appreciated.
Thanks,
Wade