This may be what I do for now. I work with computers for a living and I'm a bit of a tinkerer, so that doesn't really scare me. In the future I will probably get a more powerful pi because in the end I would like to be able to connect to it with a tablet via VNC and control it that way. I know that makes the PI do more work, but I think that is reasonable with something like a 4 or 8gb pi4.
The 4gb Pi-4B is what I'm using, with the Astroberry distro on it. Everything is done local, and I VNC into it for the user interface. Usually it's from a laptop, but in a pinch have resorted to my cell phone when the laptop battery died.
That last part of your reply brings up another question. The USB port on my camera is USB 2.0 I believe. I know that if I got a pi4 it would have a couple of USB 3.0 ports, but that wouldn't really help me at all other than future proofing. Given that the questions I have are:
1. Can I control the camera's shutter through software and still store the files on the camera's SD card?
2. Would it even matter (other than producing a little bit of read noise probably) since I usually wait a couple of seconds in between shots to let the sensor cool a little?
I ask only because the main idea was to be able to automatically do a meridian flip, plate solve to re-center the target, and then continue with the imaging sequence. I currently use an intervalometer and I could use that, but I feel like that partially makes the automation a moot point since I would still have to go stop and restart the imaging sequence. I won't be storing the files on the pi itself btw, I would use either a USB stick or my external SSD. I feel that my 64gb USB would do just fine since I almost never capture more than a gig or two of data at a time due to my camera only being 10.1 megapixels.
In theory yes, but it probably depends on the camera and what software you're running. Before getting a session manager (CCDciel in my case), I wrote a short bash script that used gphoto2 to trigger the camera. I believe there was an option to do just what you want. *Long shot*, there is a gphoto2-based DSLR INDI driver. Perhaps look there?
The meridian flip also involves stopping the autoguider, re-centering on the target, and restarting everything. The session manager should automate all that stuff too. Storing the files on the Pi isn't a big deal, as they're easily moved to the processing machine. I use the network (ftp), but it's easy enough to store the files on a flash drive and use "Sneakernet" to move them.
To the USB 2.0 camera, it will take longer to download an image, but not horribly so. Certainly won't prevent its use, just slow things down a bit. My Nikon had the same issue. 10 megapixels is relatively modest, though perfectly usable. In my case, the old Nikon really didn't like being computer controlled, so I just stayed with the Intervalometer and manually-initiated flips. Hopefully you'll have a better experience.
I recommend getting something working first, then work on optimizing it later. Here's why. The deal with the Nikon and computer control was that it affected my overall process, not just the downloading of images. For example,I was using a Bahtinov mask and the zoomed-in camera Live View for focusing, but the camera refused to let me do that once it was connected to the computer without power cycling it first. So I had to use a software-based Live View, and that didn't work very well. Very annoying. But I realized that I had to figure this out for when I would upgrade to an Astro camera, and found some focusing routines in CCDciel that would assist with this. The Nikon still had limitations that made it impractical to use them, so it was back to the Intervalometer and the mask and LCD screen. Point is, by going through that showed me what lay ahead.