One aim of the Big Amateur Telescope (BAT) is to produce high-resolution deep sky images using the Lucky Imaging (LI) approach. The low s/n ratio of LI results requires a significant amount of integration time. However, processing LI data is quite time consuming, so the idea was to stack LI results from multiple contributers to achieve s/n ratios comparable to regular deep sky imaging. To achieve this goal multiple problems have to be solved. Technical issues like the quality of the optics, collimation and focus can be solved. However, there is a quite frustrating part of deep sky lucky imaging which currently limits the LI results produced by the BAT: the Earth atmosphere. 

One misconcept of LI is that it can magically produce sub 1" resolution images, even during the worst seeing conditions. I want to illustrate the challenge of deep sky LI using a star test, that I typically carry out before a planetary/deep sky LI session to check seeing quality and collimation state of the C14. On the morning of Sep 6th 2023, I recorded a SER file consisting of  5000 subs with an exposure time of 0.15 ms per sub of the star Capella which is shown in the original version of this post. Seeing was fair to good, the Airy disc and parts of the Airy rings are almost always visible. To illustrate the difference between planetary LI and DSO LI, I integrated a certain number of subs from the SER movie, without registration or any quality rejection criterion in SIRIL, leading to integration times between 5 ms and 750 ms ...



... as you can see for typical planetary LI integration times of 5 ms - 10 ms, the Airy disc and 1st ring are well resolved. However, for DSO LI integration times between 250 ms - 750 ms, the Airy disc is already smeared over a larger part of the sky, despite the fair to good seeing conditions due to the seeing related motion of the star.

I also did some LI tests for the Fireworks Galaxy on Sep 8th 2023 with my usual C14 prime focus setup at f/7.7. Unfortunately, seeing was only average with an FWHM ~ 1.7". The usual star test on Capella after the end of the session revealed why (revision B). Seeing was poor to very poor, the Airy disc was not visible most of the time and there was no benefit of using shorter integration times, even between 5 ms - 10 ms ...



I think these examples illustrate very nicely, why DSO LI is so much more challenging when compared with planetary/lunar LI. Beside good sky transparency, you also need good seeing conditions to achieve a significant improvement in resolution. My planetary vs. DSO LI  test may be a little bit over-pessimistic, because I did not resample the planetary 0.09"/px image to the 0.29"/px sampling of the DSO LI setup. However, I think the general conclusions are still true.

Freezing the seeing below the critical 10 ms exposure times threshold would require a fast astrograph like a large RASA or Hyperstar system and/or a very bright target. I started some LI experiments with my C14 Hyperstar & ZWO ASI 183 MM for the Orion nebula. While I can reduce the exposure times below 10 ms, I was not able to catch a night with reasonable good seeing conditions to really see any benefit of this approach. Another problem with this setup is the pixel scale of 0.73"/px, which is too low for a reasonable sampling of sub-1" features. So either a camera with a smaller pixel size  or a larger astrograph is required.

Despite the issues described above, I think the Deep Sky LI approach is quite interesting, especially for targets with high surface brightness, like globular clusters and planetary nebulae.

- CS (& good seeing conditions) Daniel