"Lucky imaging" for DSO - Forums

Topic starter

Apr 13, 2023

I'm curious about "Lucky imaging" technique for DSO. Anyone tried it already?
"Lucky imaging" is often used for planetary photo, but it gets increasingly popular for DSO. This means a large number of short exposures 5-10 sec for LRGB and 10-30 sec for NB.
I wonder which camera settings to use?
- e.g. on QHY268color I normally use readout mode 0 ("DSO mode"), gain 35 and offset 5 - but for "lucky imaging" these settings need to be changed quite radically. The goal is to achieve balance between sensor's read-noise and SNR from your object to be photographed. Gain should be much higher, and possibly better to use some other readout mode.

***
Any ideas?

Helpful Insightful Respectful Engaging

andrea tasselli

Frequent contributor

Apr 14, 2023

Amateur-wise I must have been one of the first to employ that technique. Back then with CCDs you were stuck with unit gain (or whatever the mfg set) and that was it. With a CMOS sensor like the 268C I'd go with the lowest gain setting that give me the best compromise between read noise and dynamic range. I'd wager High Gain Mode (2) with gain=60 would do the trick.

Helpful Concise

David Russell

Apr 14, 2023

there is an imager named Smiller on CN who uses a 12 inch dob, and shoots 6 second exposures. his background is planetary and Lucky imaging. he wrote a PDF downloadable document that details his work flow and method for Deep Sky. very detailed. he gets some pretty acceptable results too. I am new to ABin so dont know if its acceptable to link PDF files in a thread, but PM me if you want with an email address and I will forward the PDF to you. hope that helps. Dave

Helpful Supportive

GalacticRAVE

Apr 14, 2023

e.g. @Carsten Dosche has published a number of Lucky Imaging deep sky images using an ANDOR very low readnoise CCD. Readnoise is the main problem here (besides the huge amount of data produced): you want your image to be swamped by sky, ie the sky should dominated over readnoise. In that limit, the integrated signal of an image stack is basically identical to that of one long exposure. For most practical circumstances this results in exposure times of several 10s of seconds per sub with modern CMOS. If you can push it below ~5s it would be interesting as you getting the first turbulence layers out, though you unlikely will reach the regime of planetary imaging where you have sub sec exposure times and below. but even 5 sec in most circumstances would require very low readnoise CMOS, considerably less than 1e- . I am not aware that something like that will be available at the consumer market in the foreseeable future. But for some objects/telescope combinations it may be worth a try (and has been done), e.g. in order to resolve the bright cores of objects, e.g. the planetary nebula NGC 6543.

Helpful

Nlawrie94

Apr 14, 2023 · Edited: Apr 14, 2023

Yes, members of the astro biscuit discord server have been using this technique for I think a couple years now?

One of the servers collaborative efforts made Astrobin IOTD while utilising this technique
https://www.astrobin.com/7ez0pl/

They also have other images published here on astrobin that use the technique, I like this one https://www.astrobin.com/hhegwl/
It also has a write up in the description with great info about the technique.

If you're interested in applying lucking imaging techniques I'd highly recommend looking at their website that is linked on those photos, and better yet head over to the discord server as there is a treasure trove of info regarding using Lucky imaging on deep sky objects.

Helpful Supportive

Nlawrie94

Apr 14, 2023 · Edited: Apr 14, 2023

Generally speaking, and take this with a grain of salt because I don't practice the technique, yet, but it's not worth it when you have a smaller aperture scope (below 8"), don't have a modern CMOS camera, are unable to take regularly image with exposures below 3" FWHM, dont have perfect collimation, dont have a lot of hard drive space and so on.

You need A LOT of data, if you want to beat seeing conditions, moments of calm are short and fleeting, so lots of data gets deleted and only the best kept. So lot's of HDD space, unconventional software use and so on.

If you can do the above, IIRC some members were getting stacks of exposures around the 1.4" FWHM range and then used them with regular imaging to sharpen the cores of bright objects such as galaxies.

Since it's hard to get the data, people have been collaborating on that discord server which is pretty awesome.

Jeffbax Velocicaptor

IOTD Staff

Apr 14, 2023

Hello Michael. Have a look hère

https://www.astrobin.com/groups/92/short-exposures-dso/

JF

Nauris.de

Apr 19, 2023

For some more Lucky-Imaging results you can take a look at my Gallery: https://www.astrobin.com/users/Nauris.de/

All BW-Images had been taken with a 24" Dob with a Gain of 350 and sub-length around 0.5-1.4secs.
All newer Images (color) (Veil, Horsehead, NGC 1333 and all future results) are obtained with a fast Imaging-Newt on a German-mount with 10sec subs and Gain 400-600.

So yes, it is working very nicely and in my opinion, this is a viable technique to achieve the best possible resolution. And with "short-exposures" like my 10sec examples one even can work with background-limited results.

I don't want to change back to longer subs ;)

CS,
Daniel

Helpful Engaging Supportive

GalacticRAVE

Apr 20, 2023

Very impressive results, Daniel!

Matthias

Jure Menart

Apr 20, 2023

As it's related to this thread I wanted to share this nice image from @Massimo Di Fusco : https://www.astrobin.com/jopzvq/

It really nicely shows the benefits of lucky imaging for DSO

Massimo Di Fusco

Apr 20, 2023

Jure Menart:
As it's related to this thread I wanted to share this nice image from @Massimo Di Fusco : https://www.astrobin.com/jopzvq/

It really nicely shows the benefits of lucky imaging for DSO

*** Thank you very much for sharing Jure! ***

Michele Campini

Apr 20, 2023

What about dithering in lucky imaging ?

Mau_Bard

Apr 20, 2023 · Edited: Apr 20, 2023

e.g. @Carsten Dosche has published a number of Lucky Imaging deep sky images using an ANDOR very low readnoise CCD. Readnoise is the main problem here (besides the huge amount of data produced): you want your image to be swamped by sky, ie the sky should dominated over readnoise. In that limit, the integrated signal of an image stack is basically identical to that of one long exposure. For most practical circumstances this results in exposure times of several 10s of seconds per sub with modern CMOS. If you can push it below ~5s it would be interesting as you getting the first turbulence layers out, though you unlikely will reach the regime of planetary imaging where you have sub sec exposure times and below. but even 5 sec in most circumstances would require very low readnoise CMOS, considerably less than 1e- . I am not aware that something like that will be available at the consumer market in the foreseeable future. But for some objects/telescope combinations it may be worth a try (and has been done), e.g. in order to resolve the bright cores of objects, e.g. the planetary nebula NGC 6543.

Definitively the read noise is a limiting factor. On the positive side Lucky Imaging improves greatly the seeing effect, relevant when you want to get precise details.

My two cent is that lucky imaging election field for DSOs includes detailed and bright objects where the read noise is not a problem: for instance the M42 trapezium or the Cats Eye Nebula core. In any case short exposures have to be used on these targets, to avoid pixel saturation.

Thanks for opening this interesting discussion thread!
CS to all the friends here,

Mau

Respectful Supportive

Victor Van Puyenbroeck

Apr 20, 2023

Here are a couple more users with excellent DSO lucky imaging techniques:

@Romain Chauvet
@Exaxe
@Łukasz Sujka

In general, they appear to use this technique mostly for imaging bright objects like planetary nebula or high surface brightness galaxies with large aperture telescopes. Exposure times can be 1 second or less with modern sensors.

atlejq

Apr 20, 2023 · Edited: Apr 20, 2023

If you can push it below ~5s it would be interesting as you getting the first turbulence layers out, though you unlikely will reach the regime of planetary imaging where you have sub sec exposure times and below. but even 5 sec in most circumstances would require very low readnoise CMOS, considerably less than 1e- . I am not aware that something like that will be available at the consumer market in the foreseeable future. But for some objects/telescope combinations it may be worth a try (and has been done), e.g. in order to resolve the bright cores of objects, e.g. the planetary nebula NGC 6543.

IMX462 based mono-cameras from ZWO and QHY have a claimed read noise of 0.5e-.

Filippo Scopelliti

Apr 21, 2023

Hi Michael! I show you sone images with 1 sec exposures. I hope it is useful.

https://www.astrobin.com/users/filiscoop/