Automatic removal of bad frames (on-the-fly)

[Xavery Meloman]

Hi guys.

I use 10 inch Newt with small sensors (IMX 224 size) to image small targets (M27 is the largest of them).
To deal with seeing problems I use lucky imaging, 250ms-1s exposures with high gain.
I used to record sequence of 2000-3000 frames, selected best 10% in Astrosurface, then stacked.
Apart from walking noise (no dithering) results were quite good.
Lots of files, lots of disk space, lots of time to select good ones...

Would be nice to asses FWHM on-the-fly and immediately delete bad frames, or better - put in memory buffer, asses FWHM and make decision - write file to disk or not.

I know Sharpcap has a feature - FWHM filter for live stacking. As I understand it records all the frames, but stacks only the good ones (below given FWHM). But it still keeps the bad frames anyway, right ? Or am I wrong ?

I hope guys you get my point. Is it doable in Sharpcap ?

greetings
XM

[admin]

Hi,

I see what you are aiming to do - this can be achieved in live stacking, since there is an option to choose which raw image files live stacking will save...

Set up the FWHM filtering in live stacking (you might have to be quite aggressive with the star detection settings to get decent measurements on such short exposures - while you are adjusting, turn on the option to highlight the stars so you can see that you are really getting stars picked up and measured, not noise, hot pixels, galaxies, etc).

Change the 'Raw Frames' option in the left panel of the live stacking area to 'Save Stacked' - that will make SharpCap only save the frames that are added to the stack, so the ones that fail the FWHM filter will not get saved.

clear skies!

Robin

[Xavery Meloman]

Thanks Robin, will try it ASAP.
As I undestand the same thing can be done with brightness filter (clouds/no clouds) ?

greetings & clear skies
XM

[cyrille.de.brebisson]

Hello,

I confirm that it works well, both for brightness and FWMH as I use it extensively!
Too bad this can not be also done in script when you say "capture N frames". it would be nice to be able to say: capture N GOOD frames

Cyrille

[Xavery Meloman]

Seems this method has a great potential. Here's M27, 1000x1s, Ha filter, 585 mono cam.

greetings
XM

[admin]

Hi,

glad to hear that it works Nice image! Note that for such short exposures on a deep sky subject, you are probably ending up with a noisier final stacked image than you would with an equivalent total exposure time using longer sub-exposures. However, the lucky imaging approach gives you potentially improved sharpness of the image as a trade-off...

If you want to make this happen from the sequencer and count 'good' frames then you can use the 'Live Stack for <X> frames and then save as <File Type> counting <type> frames'. Set the 'type' of frames to count to be 'Stacked' and you will have pretty much what you need

cheers,

Robin

[Xavery Meloman]

Thanks.

For now I have to deal with walking noise - it's hard to dither every 2 1s frames...
I even thought about implementing XY-movable sensor plate...
Or rotate the whole camera by random angle...
For now I dither every 5th frame with 3s settle time and don't care about rejected frames.

BTW, Sharpcap makes big eyes when I request to calculate optimal gain and offset for exposures between 0,5 and 1s

greetings
XM