Hi,
you can still end up with a small part of the trail left when sigma clipping, even if it is working largely as designed. If the trail is multiple pixels wide then it's possible to get edge pixels that are only just a bit brighter than normal - those can pass the sigma threshold criteria and still get added to the stack, leaving some visibility of the trail but much reduced from full brightness.
Still, I think that with the improvements I have planned, sigma clipping will be much improved.
cheers,
Robin
Eliminate subs with satellite trails?
-
admin
- Site Admin
- Posts: 17381
- Joined: Sat Feb 11, 2017 3:52 pm
- Location: Vale of the White Horse, UK
- Contact:
Re: Eliminate subs with satellite trails?
Hi folks,
I have just uploaded a new beta version (4.2.14920) that contains some significant improvements to sigma clipping...
From the user point of view, the biggest noticeable change is that the specification of the minimum sigma value is now in terms of ADU, not a percentage of full camera brightness range - that gives much finer control around the areas that are of most interest (typically zero to maybe 16 ADU). Note that for cameras operating in high bit depth modes, the ADU specified here are always 16 bit ADU values, even if your camera is perhaps only working at 12 bits - in which case 16 ADU at 16 bit is equivalent to 1 ADU at 12 bit... It may be possible to get good results setting this to zero - if you see pixels that seem 'stuck' in the stack and don't change then turn this up a bit to give some more flexibility for SharpCap to include values further away from the current mean.
From the results point of view, the big change is that SharpCap now tracks the maximum and minimum value at each pixel position in the image during the initial establishment phase (the initial 10 frames - or however many you have selected). At the end of that phase, SharpCap checks to see if the minimum/maximum values recorded for each pixel are anomalous and throws them out if they are. This will remove satellite trails from the stack when the initial phase ends without the need to 'unstack' problem frames - the trails will show initially and then vanish when the frame count of the stack hits the 'initial frame count' value. This only works perfectly if there is at most one bright and one dark anomalous value at each pixel position in the image - very likely fine for satellite trails, but less likely for hot/warm pixels that might stay roughly stable in position during the first few frames.
To counter the problem of hot/warm pixels leading to brighter values in the establishment phase that then get 'locked in' because the normal pixel values in that part of the image are considered 'out of range', SharpCap will now let a pixel value through even if it is outside the normal sigma limits if the stack has rejected the value at the same pixel position for the last 10 frames - that helps let 'normal' pixel values start to reduce the brightness of those locked in hot pixels, which then improves the overall image. I'm not sure if this value of 10 should be user-adjustable in the future.
cheers,
Robin
I have just uploaded a new beta version (4.2.14920) that contains some significant improvements to sigma clipping...
From the user point of view, the biggest noticeable change is that the specification of the minimum sigma value is now in terms of ADU, not a percentage of full camera brightness range - that gives much finer control around the areas that are of most interest (typically zero to maybe 16 ADU). Note that for cameras operating in high bit depth modes, the ADU specified here are always 16 bit ADU values, even if your camera is perhaps only working at 12 bits - in which case 16 ADU at 16 bit is equivalent to 1 ADU at 12 bit... It may be possible to get good results setting this to zero - if you see pixels that seem 'stuck' in the stack and don't change then turn this up a bit to give some more flexibility for SharpCap to include values further away from the current mean.
From the results point of view, the big change is that SharpCap now tracks the maximum and minimum value at each pixel position in the image during the initial establishment phase (the initial 10 frames - or however many you have selected). At the end of that phase, SharpCap checks to see if the minimum/maximum values recorded for each pixel are anomalous and throws them out if they are. This will remove satellite trails from the stack when the initial phase ends without the need to 'unstack' problem frames - the trails will show initially and then vanish when the frame count of the stack hits the 'initial frame count' value. This only works perfectly if there is at most one bright and one dark anomalous value at each pixel position in the image - very likely fine for satellite trails, but less likely for hot/warm pixels that might stay roughly stable in position during the first few frames.
To counter the problem of hot/warm pixels leading to brighter values in the establishment phase that then get 'locked in' because the normal pixel values in that part of the image are considered 'out of range', SharpCap will now let a pixel value through even if it is outside the normal sigma limits if the stack has rejected the value at the same pixel position for the last 10 frames - that helps let 'normal' pixel values start to reduce the brightness of those locked in hot pixels, which then improves the overall image. I'm not sure if this value of 10 should be user-adjustable in the future.
cheers,
Robin
-
carlomuccini
- Posts: 257
- Joined: Mon Apr 27, 2020 12:42 pm
- Location: Montecatini Terme (PT), Italy
- Contact:
Re: Eliminate subs with satellite trails?
wow !
definitely worth trying as soon as possible
Carlo
definitely worth trying as soon as possible
Carlo