Re:Automating the capture of airplane pictures with Raspberr...
by ShakataGaNai on Thursday May 11, 2023 @11:33AM
(#35903634)
Most of this project is "Raspberry PI and ADSB" which has been done before, the interesting part is taking pictures. Unfortunately, this solution is very niche. If you happen to live in a location where planes only fly in a very specific location, you can take a picture. It's cool for a one-off solution, but probably not something most people can use.
https://skybot.cam/about is a little more generically interesting as it takes a camera and points it at the aircraft. Then some vision detection to verify that a plane was actually captured, then post it.
Re:Automating the capture of airplane pictures with Raspberr...
by tpmx on Thursday May 11, 2023 @11:30AM
(#35903584)
Does anyone know of any (perhaps open source) hardware that's suitable to aim a DSLR-sized camera in a suitable direction with a reasonable kind of precision? I guess medium power (around 10-20W?) BLDC motors combined with an hall effect sensor would be suitable for this application since they are used in gimbals?
It would be neat to have an auto-aiming and tracking camera with a telephoto lens and record a close-up video of the aircraft instead of a wide angle photo.
Re:Automating the capture of airplane pictures with Raspberr...
by dekhn on Thursday May 11, 2023 @11:53AM
(#35903974)
I've done this but for observing the sun with a DSLR/telephoto for long periods. I've seen several projects to do this, a related project is https://opensamofficial.github.io/OpenSAM/ but that's more of a stabilizer.
When I built my own I hit a number of challenges: the 3d printed plastic parts melted in the sun (switched from PLA to PETG, printed in pure white). DSLRs with telephotos are heavy so I ended up using gearing. I should have put more time into balancing the camera to reduce the motor demands. I don't think a DSLR is ideal- I'd much rather use a small, cheap industrial camera, but I've never found the equivalent of an affordable 300mm telephoto that is C-mount and goes on industrial camera. I used stepper motors, not BLDC, without feedback (other than a zero-position limit switch) and a simple realtime algorithm for pointing; it worked, but wasn't really good enough to make the sunspot timelapses I wanted.
I think for this it makes more sense to buy a quality telescope motorized mount, a nice refractor telescope, and then put a DSLR at prime focus. It'd be about $3K but would save a lot of time compared to DIY.
Re:Automating the capture of airplane pictures with Raspberr...
by tpmx on Thursday May 11, 2023 @12:06PM
(#35904183)
I didn't think about the heat. Yeah, a DSLR with a decent lens is quite heavy. And expensive, so it needs to be protected.
It's a tricky problem to do on the cheap.
I don't think I'm prepared to spend $3K, my idea was rather to spend some fun time on HW work as a diversion from the regular paid SW work - and to put some use to that expensive camera gear I'm not really using.
Thanks for your lessons learned; they have been helpful!
Re:Automating the capture of airplane pictures with Raspberr...
by gavinsyancey on Thursday May 11, 2023 @12:57PM
(#35904920)
Might be worth looking into motorized telescope mounts. Should have plenty of precision, although you might not be able to slew very fast and will probably have to handle zoom/focus/camera control separately. Looks like people have built and documented open-source telescope mounts, or you can buy a cheap one for a couple hundred dollars.
Re:Automating the capture of airplane pictures with Raspberr...
by MandieD on Thursday May 11, 2023 @02:37PM
(#35906362)
The steerable rotors amateur radio operators use with large directional antennas might be of interest. The EME (Earth-Moon-Earth) and radio satellite communities in particular have many automated antenna rotation solutions.
https://skybot.cam/about is a little more generically interesting as it takes a camera and points it at the aircraft. Then some vision detection to verify that a plane was actually captured, then post it.
It would be neat to have an auto-aiming and tracking camera with a telephoto lens and record a close-up video of the aircraft instead of a wide angle photo.
When I built my own I hit a number of challenges: the 3d printed plastic parts melted in the sun (switched from PLA to PETG, printed in pure white). DSLRs with telephotos are heavy so I ended up using gearing. I should have put more time into balancing the camera to reduce the motor demands. I don't think a DSLR is ideal- I'd much rather use a small, cheap industrial camera, but I've never found the equivalent of an affordable 300mm telephoto that is C-mount and goes on industrial camera. I used stepper motors, not BLDC, without feedback (other than a zero-position limit switch) and a simple realtime algorithm for pointing; it worked, but wasn't really good enough to make the sunspot timelapses I wanted.
I think for this it makes more sense to buy a quality telescope motorized mount, a nice refractor telescope, and then put a DSLR at prime focus. It'd be about $3K but would save a lot of time compared to DIY.
It's a tricky problem to do on the cheap.
I don't think I'm prepared to spend $3K, my idea was rather to spend some fun time on HW work as a diversion from the regular paid SW work - and to put some use to that expensive camera gear I'm not really using.
Thanks for your lessons learned; they have been helpful!
That's what I did. This time around, I'd just buy that.