2
$\begingroup$

I remember that there are five types of nebulae:

  1. Planetary nebulae have not much to do with planets, but are emission nebula consisting of an expanding, glowing shell of ionized gas ejected from red giant stars late in their lives.
  2. H II regions are emission nebulae in which interstellar atomic hydrogen is ionized.
  3. Reflection nebulae are "clouds of interstellar dust which might reflect the light of a nearby star(s)".
  4. Dark nebulae are interstellar clouds that is so dense that it obscures the visible wavelengths of light from objects behind it.
  5. Supernova remnants are just that - the leftovers of a supernova

As Donald Knuth once said

Science is what we understand well enough to explain to a computer. Art is everything else we do.

I fully understand the definitions of the different types of nebula, but I struggle to write a computer-readible definition for each of nebulae types. What I mean: Given a deep-sky image with pre-identified nebulae zones, I do not know how a nebula type identification algorithm would look like. Do different types of spectra uniquely define each nebula type? That would work for the emission nebulae at least, I guess.

References

The answer is possibly given by Steven R. Coe's article What Are All These Different Types of Nebulae, and What Details Can I See in Them with My Telescope? which hidden behind a paywall and not accessible for me.

Improve this question
$\endgroup$
6
  • 1
    $\begingroup$ Knuth doesn't seem to be aware of computer vision in that quotation... We can't even tell a computer what a cat is, in terms that allow it to tell from an image whether there is one on there or not - not without resorting to machine learning (which is what you do when you've given up on finding a definition that can be programmed). $\endgroup$
    – Kristoffer Sjöö
    Commented Jan 16, 2021 at 21:45
  • $\begingroup$ Well, I purposedly did not mention AI and the fact that this is the usual approach of doing such classifications. Assuming that some machine learning algorithm already attributed a type to a given nebula zone on an image, one could in a 2nd step plot a state space spanned by different variables (e.g. mean wavelength, elements identified in the spectrum etc.) and analyze whether one find some semi-hard criteria. $\endgroup$
    – B--rian
    Commented Jan 16, 2021 at 22:26
  • $\begingroup$ Differently put: Adanced feature selection methods would allow you to determine the hard criteria I am after. $\endgroup$
    – B--rian
    Commented Jan 16, 2021 at 22:34
  • 1
    $\begingroup$ Planetary nebulae emit light in the doubly ionized oxygen (O III) line. I know this is a way to tell them apart from stars when they have a very small angular diameter. Not sure about the other types of nebulae—maybe H-alpha for H II regions? $\endgroup$
    – Pierre Paquette
    Commented Jan 16, 2021 at 22:49
  • 1
    $\begingroup$ @B--rian Sorry it looked really badly :-) I also voted the post up. My goal was actually this tag wiki edit suggestion review: OP stated that the "untagged" tag is mod-only, I also tested it (if I can retag it as "untagged", then I had to roll it back and then reject the edit suggestion, however an error message said that yes it is really mod-only). $\endgroup$
    – peterh
    Commented Jan 17, 2021 at 0:51

0

Reset to default

You must log in to answer this question.

Browse other questions tagged .