M57. Working to improve image sharpness

A place to share images that you have taken with SharpCap.
Forum rules
Please upload large images to photo sharing sites (flickr, etc) rather than trying to upload them as forum attachments.

Please share the equipment used and if possible camera settings to help others.
Post Reply
timh
Posts: 242
Joined: Mon Aug 26, 2019 5:50 pm

M57. Working to improve image sharpness

Post by timh »

An update on trying to get 'lucky imaging' to deliver improved resolution. Here under the Heathrow flight path sky seeing never gets better than FWHM ~ 2.5-3 arc sec.

Image 1 is an integration of 183 x 5s exposures at gain 200 - carefully selected from about 270 to eliminate those with FWHM > 3.3 and high Eccentricity etc. It represents about the best I typically get using conventional methods.

SW Synscan 250PX Dobsonian with 0.9X SW flattener reducer (F 4.16 and f = 1050mm). SW electric focuser.
ZWO AS1294MC PRO camera BIN1 with 4.3 um pixels and 14 bit ADU. Image scale 0.9 arcsec/ pixel at -10C.

The idea here was to improve on this image by transferring the luminance from a higher resolution image obtained by lucky imaging.

Pilot observations of the double double Epsilon Lyra were first carried out in order to scope out suitable conditions for lucky imaging. The two star pairs, E1 and E2 are separated by 2.3 and 2.4 arcsec respectively.

Images 2a and 2b are individual 'good' frames chosen from about 200 of the E2 pair. 2a is a 5ms frame sampled at an image scale of 0.9 arcsec pixel and 2b the same but at an image scale of 0.45. About 5% of the frames were of similar quality Not surprisingly the finer 0.45 scale gives the better resolution.

Image 2c is an example of a 'good' 300ms frame sampled at 0.45 arcsec/pixel and Image 2d of a 3s frame. Both are much dimmer than the 5ms frames because I had to cut down the light by interposing an SII filter in order in order to avoid saturating the 12 bit ADU. Surprisingly perhaps - given that the time is so much longer than air turbulance changes - the 300ms frame still shows quite good resolution and even the 3s frame still shows some enhanced resolution.

For lucky imaging of the ring nebula the conditions were as follows

i) Images collected by Sharpcap as ~ 15 min .SER files
ii) Exposure time 300ms and gain 285
iii) Frame size 1800 x 1200 pixels (810 x 540 arcsec) - must be big enough to contain sufficient starts for alignment
v) Sampled at 0.45 arcsec/ pixel

SW Synscan 250PX Dobsonian with 0.9X SW flattener reducer (F 4.16 and f = 1050mm). SW electric focuser.
ZWO AS1294MM camera in BIN1 46Mb mode with unbinned 2.315 um pixels and 12 bit ADU (0.45 arcsec/ pixel)

Rejection and selection of frames for stacking is a critical part of lucky imaging. SER files were converted to thousands of FIT files using the PIPP freeware. The PIPP contrast-based quality algorithm was used to select the top 70%. Then subframe selector in PixInsight was used to select only those frames showing more than a certain number of stars (bright enough) that were well defined (FWHM < ~ 1.7 and Eccentricity < 0.65). Pass rates varied from night to night and within nights between as low as zero up to as high as 10%. Visual inspection using BLINK may be superior but is laborious for large numbers of frames.

Image 3 shows the result of stacking about 2000 x 0.3s -(total ~ 10min) - of selected frames.

Image 4 shows the final colour result obtained by replacing the luminance of image 1 with that of of image 3

The overall process looks to have substantially improved the resolution fron an FWHM of about 3.2 arc sec to about half of that. This is illustrated by the successful resolution of a (line of sight) triple star close to M57 with components of differing brightness that are all within 2 arc sec of each other (see image 3).

Undoubtedly resolution could be improved with shorter exposures. However this would come at the cost of more noise (including accumulated raed noise) and for each target object a careful compromise needs be struck between resolution and achieving a workable level of quality in each frame.


TimH
Attachments
M57 image4.JPG
M57 image4.JPG (35.61 KiB) Viewed 196 times
M57_image3.JPG
M57_image3.JPG (14.37 KiB) Viewed 196 times
E_Lyra_exposuretimes.JPG
E_Lyra_exposuretimes.JPG (18.62 KiB) Viewed 196 times
M57_090621_183x5sframes.JPG
M57_090621_183x5sframes.JPG (12.44 KiB) Viewed 196 times
User avatar
admin
Site Admin
Posts: 6964
Joined: Sat Feb 11, 2017 3:52 pm
Location: Vale of the White Horse, UK
Contact:

Re: M57. Working to improve image sharpness

Post by admin »

Hi,

what an interesting experiment and write up - I loaded the two M57 images separately and enlarged 'Image1' until the size matched 'Image4', which made it quite clear how much more detail you have managed to bring in!

I wonder if you could (with appropriate software) use double stars for frame quality selection? If you sample pixel brightnesses from a line joining the star centers then good frames would have a lower minimum value in the middle than bad frames. In one way this is just a different measurement of the FWHM, but FWHM measurements on small, noisy stars can be inaccurate, so maybe this would have value.

cheers,

Robin
timh
Posts: 242
Joined: Mon Aug 26, 2019 5:50 pm

Re: M57. Working to improve image sharpness

Post by timh »

Thanks Robin,

It was certainly an interesting journey finding out how to make this work. With fast optics and ever better low read noise CMOS cameras maybe 'lucky imaging' will become more mainstream for a wider range of targets ?

It can be very 'clunky' though both around data storage and, as you point out, then the problem of post selection from thousands of frames. In addition I am manually steering the ROI selection window as the SER file is captured because the Dobsonian AZ tracking is erratic --and currently I can't control it from within SC (but am working on that - needs a diiferent serial to USB cable). This would be much better because then I could dither and potentially be able to use brighter frames at BIN2 and drizzle for similar resolution?

Maybe there is an opportunity to have SC drive this type of astrophotography too in the future in ways that would take the pain out of it?

The ultimate I think would be to have livestacking incorporate frame selection on the fly and perhaps in the kind of way that you suggest (Could some sort of global evaluation of the "depth of valley" between say all star pairs within a certain range of spacing be calculated fast enough on the fly ?)

Thinking about it I may also have missed a trick that could already be tried with Sharpcap as it stands ? Maybe simply livestacking say 0.5s FIT frames and applying the brightness and FWHM filters is itself a 'weak' lucky imaging strategy that already could provide a smaller but useful improvement in resolution but in a painless way --- or at least provide a useful initial selection step ?

TimH
User avatar
admin
Site Admin
Posts: 6964
Joined: Sat Feb 11, 2017 3:52 pm
Location: Vale of the White Horse, UK
Contact:

Re: M57. Working to improve image sharpness

Post by admin »

Hi Tim,

it would certainly be worth a try with the FWHM filter to see if that helps - set the target quite low to include only ~50% of frames. You can rig up the RAW frame saving to save everything just in case you need to revisit the selection later.

I also wonder wether feature tracking in 4.0 might be able to use the ROI to keep a target like this centered? 4.0 now has the capability to track using the ROI rather than the mount and also a 'center of brightness' tracking mode that might work better with faint stuff than the image feature detection.

cheers,

Robin
timh
Posts: 242
Joined: Mon Aug 26, 2019 5:50 pm

Re: M57. Working to improve image sharpness

Post by timh »

The feature tracking is an excellent idea --thankyou. I''ll try that along with short FWHM_selected subs in livestacking next time out!
Post Reply