Zero-In

Introduction

Zero-in is a proof-of-concept for an advanced plate-solving push-to system for amateur telescopes that I wrote as a personal project for my own use.

The software estimates the position of the telescope by combining information from a camera looking at the sky and inertial motion sensors (similar to the ones used in all commercial smartphones today). The software is designed to work with or without an equatorial platform and with no mechanical alignment of finder scope or motion sensor required.

This is not a release, it is a proof-of-concept.

Because this is a proof-of-concept, it requires significant technical skill to set up and use this system. As of this writing, my intention is to provide this as a reference for the novel alignment and calibraton algorithms implemented here as well as a proof-of-concept for what a system focused on catering to the needs of advanced deep-sky observers might look like.

What is plate-solving?

Plate-solving refers to the idea of comparing a photograph of the sky against a database of known patterns of stars to identify a mapping between pixels and positions on the celestial sphere (i.e. celestial coordinates). A landmark paper in this domain was associated with the creation of astrometry.net by Dustin Lang et. al. This project uses the binary form of astrometry.net engine. Please see the rpi_setup_readme.org file for how to set up a Raspberry Pi appropriately with astrometry.net in the correct configuration

What hardware do I need?

1. An electronic finder scope, consisting of a camera and a lens to image the sky.

2. A 3-axis motion sensor that can output a quaternion describing its orientation.

3. A way to mount these on the telescope, with the finder scope pointing roughly in the direction of the telescope

4. A computer. I have gotten it working on a Raspberry Pi but was unhappy with the stability. I just settled to pulling USB cables to my laptop. It's messy but it works fine and lets me get on with deep-sky observing.

For the electronic finder, I specifically used a ZWO ASI 290MC-Cool camera (way overkill) coupled to an Orion 70mm multi-use finder scope (also overkill).

For the motion sensor, I used an Arduino Uno board with an Arduino 9-axis motion sensor shield built around BNO055 (with the standard libraries that come with it).

What is the advantage of plate solving?

It is independent of mechanical errors in the construction of the telescope. The only things that can affect it is the relative flexure of the electronic finder scope with respect to the main telescope and the temperature drift. This way, complex pointing models are not needed.

How do you avoid all mechanical alignments?

1. For the finder-scope to telescope alignment, a single-star calibration is needed. I center a star in my telescope, and simply tell the software which star I have centered. The alignment process takes a plate solve and stores the offset RA/Dec from the plate center to the star, and this is good enough. I only need to "re-align" when I recollimate, and this is even simpler because of the interface.

2. Motion sensor calibration is automatic. Every time the system gets appropriate plate-solve results, it will automatically recalibrate the motion sensor orientation. I designed this so that you can literally duct tape the motion sensor to the telescope -- if the tape gives, you can just tape the motion sensor back and the system will recalibrate within a few plate-solves!

Does it handle an equatorial platform / untracked / alt-az tracked scope?

Yes. It handles all of these.

The coordinate mathematics are all in the code. To handle an equatorial platform, the code recognizes that:

1. The telescope's alt/az axes are frozen in an alt/az frame whose LST is equal to the LST when the platform becomes flat.

2. The motion sensor has a sense of true altitude from the accelerometer's sense of gravity.

So with appropriate coordinate transforms, an equatorial platform is easily handled -- the software only needs to know the travel time of the platform and needs to be informed every time the platform is reset.

Okay, walk me through the software stack

The UI is written using the Qt Framework through PyQt5. Communications with the camera hardware is handled using INDI via PyINDI. Some coordinate math is implemented by hand where simplicity and speed are of the essence. Others are sourced from astropy. Python libraries are used heavily wherever available to simplify the implementation. The plate-solving itself is done by invoking astrometry.net's solve-field binary, but the source extraction is done using SEP.

The system is able to display a correctly oriented DSS image as long as a webserver is available to serve the NGS-POSS data. Patrick Chevalley of SkyChart has adapted the NGS-POSS code to Linux and I have written a small web-server for this purpose around it, which I shall release separately.

main.py is the entry-point for the program. Most of the action happens in backend.py which handles various tasks including invoking the solver and handling some math.

Alt-Az Workshop Presentation

I presented the general ideas behind this proof-of-concept (along with a crude demo video) at the 2021 Portland Alt-Az Workshop. The video recording of the talk is available here and the slides are available here

White paper

A white paper on the IMU recalibration algorithm is expected eventually

Copyright and License

The code (where not sourced from other projects) is Copyright (c) Akarsh Simha 2023. My code is licensed and released under GNU GPLv3.