I've been developing a new automation suite for my imaging based on Python and last night was a hazy night with a bright Moon so I just used it as an engineering night to test the code. Set it up to image as many Messier globular clusters and it worked its way through these. Lost several targets due to clouds which gave me a chance to work on and think about how to resume sequences interrupted by weather.

Each image is RGB 15x15s each filter. No flats and quickly stacked and stretched in PI. My intention wasn't pretty pictures but to exercise the software and verify its operation. Still it is neat to see the relative size and density of some common globular clusters seen together like this.

I've tried to abstract the device layer so the same code will run on ASCOM or INDI and so far it seems to be an attainable goal.

Focusing is using a method similar to Focusmax where you determine slope parameters for a given imaging configuration and then use a bright star and focus from one side of best focus and move in the same direction for the entire focus operation. I find this preferable as it avoids backlash issues. I also have the code find a bright nearby star to focus on which I find preferable for narrow band filters as you can use very short quick exposure times and get high enough SNR to resolve the defocused star.

Slews can be improved using PlateSolve2 under ASCOM/Windows or astrometry.net (local) for INDI systems. I also have a fallback to nova.astrometry.net online.

My motivations were to improve my knowledge of Python 3 and I wanted a framework to let me build a system to create a list of targets and have them automatically imaged while I slept. There are commercial solutions to do this but they are either expensive or had design decisions I wasn't happy with so I figured trying to do it myself would be a fun experience.

Code is still in a very early state but if I get something I think would be interesting to other people I plan on making it all available. I want to try to get it working reliably and well documented so it may be awhile.