Open source code for my plate solver system

[Akarsh Simha]

Here it is

https://github.com/kstar/zero-in

Maybe I will provide instructions to set it up when I’m free. Right now I’m not. I just wanted to get this out of the door now that I’m temporarily liberated from the encumbrances of employment.

Regards

Akarsh

[Ted Hauter]

A new way to find objects since 1980ish? Truly revolutionary. 

Don't get too used to having months off to actually get to your dreams. I did and got done what I wanted but then had to return. Luckily I lucked out upon return and can now chop away at some debt, but another dream on hold 😂 

I gotta see this in action. Can you make a video please 🙏 sounds like you have time 😇😌

--
TAC Astro-Classifieds: https://groups.google.com/forum/#!topic/sf-bay-tac/extFBoHlCe4
 
Observing Sites, Observing Reports, About TAC linked at top of:
 
https://groups.google.com/forum/#!forum/sf-bay-tac
 
Subscribers post to the mailing list at:
 
sf-ba...@googlegroups.com,
---
You received this message because you are subscribed to the Google Groups "The Astronomy Connection (TAC)" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sf-bay-tac+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sf-bay-tac/CA%2B9k5tzd0d3jdDwUwPwmjE8qX2KGP4QB_vA5htegLCPVQ8dNWQ%40mail.gmail.com.

[Akarsh Simha]

On Sat, Nov 25, 2023 at 6:39 AM Ted Hauter <thgo...@gmail.com> wrote:

A new way to find objects since 1980ish? Truly revolutionary. 

Plate solving has been around forever, but I think it has really taken Sony's innovation in fast CMOS sensors (the IMX... series), the entire maker ecosystems with Raspberry Pi / Arduino etc, and fast processors to make it work as a push-to system. Astrophotographers have been using it with go-to mounts for ages, but this is about push-to. The hardest challenge here is that the fastest plate-solve I know of (PiFinder using Tetra3) takes 0.5 seconds. Waiting 0.5 seconds to know how much you've pushed your telescope and how much you've to push is preposterous -- imagine the amount of backlash in the system. A human would probably need feedback at 10 Hz, so 2 Hz is frustrating to use. To bridge the gap, my system and the PiFinder incorporate commercial accelerometers (inertial motion units, or IMU) which are capable of tracking position at kHz rates, but are utterly imprecise outside of NASA and defense. The combination of a plate solver and the IMU is what makes a push-to possible. My innovation there is the algorithm to combine these two, that is agnostic to how the IMU drifts or how it is placed relative to the telescope. There is still lots of scope for improvement here. I don't think this was possible for hobbyists in the 1980s, but it is today.

 

Don't get too used to having months off to actually get to your dreams. I did and got done what I wanted but then had to return. Luckily I lucked out upon return and can now chop away at some debt, but another dream on hold 😂

Very valid caution. I am cognizant of the risk, but it is definitely easy to become complacent. I feel it already a week into my funemployment.

 

I gotta see this in action. Can you make a video please 🙏 sounds like you have time 😇😌

I have an old crappy video:

https://www.youtube.com/watch?v=VLSbmhEUWO0&t=8850s

Regards

Akarsh

To view this discussion on the web visit https://groups.google.com/d/msgid/sf-bay-tac/CAL-cFaEPnvwb0RR3-pzX28KJ8VkTkmrU0pWXZyn4w5mJoMxTxQ%40mail.gmail.com.

[Ted Hauter]

Was looking for the Ad but Celestrons first go-to computer telescope was in a late 80s Sky and Tel for 14k! That was the price of not one new car but two new cars! ( The Honda CRX that I wanted so bad in High School but couldn't fit in (so small a car! but 50 mpg) was $8500.)

The software. Congratulations 👏  it's a beauty!

 

To view this discussion on the web visit https://groups.google.com/d/msgid/sf-bay-tac/CA%2B9k5txVKZdipTD%2BPed1e1yhgvNTuL_Cy76vrsmQaqt7LtQGgg%40mail.gmail.com.

[Peter Santangeli]

Very cool.

 How would you differentiate this from the pifinder code?

pete

To view this discussion on the web visit https://groups.google.com/d/msgid/sf-bay-tac/CA%2B9k5txVKZdipTD%2BPed1e1yhgvNTuL_Cy76vrsmQaqt7LtQGgg%40mail.gmail.com.

[Akarsh Simha]

On Sun, Nov 26, 2023 at 19:30 Peter Santangeli <pe...@santangeli.net> wrote:

Very cool.

 How would you differentiate this from the pifinder code?

pete

Richard and I did not know about each other’s work until we met at GSSP 2022, so they are independent attempts at the same thing — one is in a crude proof-of-concept stage and another is a refined product. At GSSP, we put both our systems on my telescope and compared them. Although his work is production ready, there are things in my system (such as handling equatorial platforms) that can be ported over. My hope is that I’ll build myself a PiFinder and work to transfer my ideas to the PiFinder software. I have way too many projects so I don’t know how much time I’ll end up committing myself, but one of my prime motivations to open source this is so that anyone could contribute to this effort.

Regards

Akarsh

To view this discussion on the web visit https://groups.google.com/d/msgid/sf-bay-tac/CALeFARycxXLOOm%2BzhVoYszOX%3D1zWTBJ%2BdXCuNgJvK9-Faa0rRg%40mail.gmail.com.

[Alex]

Thank you Akarsh. I definitely plan to check it out and compare it to our work on the same pointer feature a couple years back on the Android platform  (I'm not affiliated with that code anymore though).

At the moment, I'm still playing with my recently acquired Portaball otfitting. In particular, with regards to a robust pointing system for it. The "PS with accelerometers fusing" math turned out to be an order of magnitude more complex now (add another motion dimension of the Portabal mount), errors are piling up much faster too, so I've ditched that effort for the time being. My other electronic pointers work way better for me personally :)