Looking for testers of new Polar Alignment utility.

[themos]

Hello,

I am new to this forum, I am usually found in the UK-based stargazerslounge.com

If you would like to volunteer to try out a new Polar Alignment method, please let me know. The requirements are:

You can see the Celestial Pole region from where you usually set up.

You can attach a camera to the mount and bring the JPEG image files to your computer.

You can freely rotate the RA axis (this excludes some AstroTrac mounts and similar, I think).

You have Internet connectivity (to connect to nova.astrometry.net for plate solving although other local plate-solving methods will be supported in the future, such as AstroArt).

You can run Python on your computer (e.g. Anaconda for Windows, Linux, MacOS X)

You might be able to verify your Polar Alignment within 2-3 minutes. The software will be in the public domain. It does not require a GoTo mount, it can be used to polar align even a barn-door tracker.

This is the basic idea (excuse the use of Northern Hemisphere conventions)

After a rough polar alignment, I set the camera somewhere close to declination 90 so I can see Polaris in the field of view. Keeping the camera fixed to the mount, I slew in RA by a large angle (40 degrees say) and watch where Polaris goes. If it goes off the FOV, the RA axis is not close to Polaris. So I adjust the polar alignment, bring Polaris back into the FOV and try again (the details of this iteration need to be worked out). Once it's set so that Polaris stays in FOV after a large RA slew, we are good to fine-tune the polar alignment by PhotoPolarAlign. The first two images give us the position, in pixel coordinates, of the RA axis (the red cross). I take the first image at an RA position so that the sensor is taking a picture of the sky in "portrait mode" and the second image in "landscape mode", with the long side of the sensor horizontal. The plate solving tells me where the NCP is (allowing for precession since 2000), in pixel coordinates and what the scale is. Now that the camera is horizontal, x and y pixel offsets translate directly into instructions to move the mount RA axis up-down or right-left by so many arcminutes. After I adjust the polar alignment, I take another image and check for improvement. The RA axis pixel coordinates have not changed, only the NCP pixel coordinates move. With practice, it should take only 3-4 iterations to get it close enough (whatever close enough is for your imaging needs).

The thumbnails give you a taster of what it looks like up North. A newer version will be able to deal with Southern Hemisphere.


Please download the installer (https://drive.google.com/file/d/0Bws-dmFqKZThNThQeWgtc2UxaDg/view?usp=sharing) and run it, selecting your Documents folder as the destination (any folder where you have permission to write files will do).

Before running PhotoPolarAlign, you will need to have a Python 2.7 installation that includes all the modules it needs. (I used the Anaconda distribution (https://store.continuum.io/cshop/anaconda/))

You will also need to register with http://nova.astrometry.net/ and get an "API key".

AstroPolarAlign can be started by double-clicking the PPA.py file that the installer puts on your file system.

The first thing to do is to add the key you got from nova. Use the File -> Settings menu for that.

The second thing is to try out the supplied test image files (from the Northern Hemisphere). The two big buttons to the left of the Operations panel are used to select image files. Please select the file v.jpg for the top one and the file h.jpg for the bottom one. Hit the Solve button next to the big button, first for one image, then wait until the "Solved" label goes green, then do the second one. This will need an internet connection, of course. Once both "Solved" labels are green, you can hit the Find Polar Axis button and wait for the annotated image to pop up. Your polar alignment is also displayed on the main screen.

Then select the i.jpg file with the big button on the right side of the Operations panel. Click its Solve button and wait for a solution. Then click on "Show Improvement" button to see a new plot.

Then, exit the program and restart it to try it with your own images (the reason is that the scale is remembered in a single session but not across sessions and it's unlikely that your image scale will match the examples).

Thanks for your time, hope you will find it useful.

Themos Tsikas (themos.tsikas@gmail.com)

[bkm2304]

Hi Themos,

Welcome to Ice In Space. This looks very interesting. Sadly my scope is on a pier and doesn't get to actually see the SCP due to the roll - off roof hogging that region! But I am sure there are plenty of people on Ice In Space who will give your program a go!


Cheerio and good luck,:thumbsup::welcome:

Richard

[themos]

Thanks Richard!

[themos]

Hello,

Here is an update (v1.0.2) (https://drive.google.com/file/d/0Bws-dmFqKZThSlpIdHRUS0dxRms/view?usp=sharing)for the Southern Hemisphere. I add circles for Sigma and Chi Octans so that you can verify that the plate-solving makes sense (if the circles have stars in the centres). Also, a better installer (and uninstaller) have been used.

Example:

[themos]

Hello

Here is Version 1.0.3 (https://drive.google.com/file/d/0Bws-dmFqKZThd253aUl3dWZUWW8/view?usp=sharing) with the ability to use a local astrometry.net solve.

When you first run it, the Local buttons should be disabled. You need to go to AstroTortilla and create two new configurations, called PPA_coarse.cfg and PPA_fine.cfg. You can just copy the one configuration that works for you to both these names (I use AT's File -> Load Settings... and File -> Save Settings...). If those files are present, PPA will, on restart, have the Local buttons enabled.

Personally, I created two "Backend config" files for astrometry.net's solve-field command: one with just the 4100-series indices that is good for widefield images and a second one with just the 4200-series indices. The first time you do a local solve in PPA, the PPA_coarse is used, in subsequent solves the PPA_fine is used. We only copy these AT settings:

Backend config, Downscaling, Scale minimum, Scale maximum, Scale units, Cygwin shell, Custom options.

In future, depending on feedback, we may only steal settings from one AT settings file.

[cfranks]

Hi Themos,

I will try and test it tonight. I'm having problems with my Losmandy G11 + Gemini2 and your system might help. I've never used Python nor AstroTortilla but I will let you know how I go.
Regards
Charles

[Camelopardalis]

Any new method for getting good polar alignment sounds good to me :D

Does it require view of the pole?

[themos]

Hello!, yes it does.

[themos]

Thank you Charles. I'd advise to try out the functions in the daytime, using the example images provided, just to make sure you're set up correctly.

Themos

[mswhin63]

I'll keep this in mind. I just set-up a good alignment just last night. I still have to get Internet connection to my set-up but when I do I will try it out.

What kind of computer can it run on? I have a very low processor and software (XP) solution.

[themos]

Hello Malcolm,

It doesn't need a fast computer if you're submitting images to nova.astrometry.net for plate-solving. On Windows XP, there may be an issue that I have a fix for, so message me if there's any problem.

Themos

[themos]

Hello again,

Version 1.0.4 is ready, Windows setup (https://drive.google.com/file/d/0Bws-dmFqKZThUjBtc2FCeENOdXc/view?usp=sharing) or Python file (https://drive.google.com/file/d/0Bws-dmFqKZThNm9aZDBoVmI4c3M/view?usp=sharing)

Pathname blanks should be handled correctly,
dd:mm:ss display,
detects wrong parity (mirrors in optical train) and incompatible image dimensions,
local solver settings dialog,
local solver settings maintained in own settings file,
can import local solver settings from any AstroTortilla configuration file

[shamantanthew]

Hi Themos, I'd love to give your software a shot. I hope you're still around as I realize I'm digging up a rather old thread! I'm wondering if you can elaborate on the above for 1.0.4. I've installed AstroTortilla but I'm not familiar with it. How do I go about configuring the local files for your software? I'm using a Canon 6D and 200mm f/2.8 and 400mm f/5.6 lens, so nothing super narrow as far as field of view is concerned. I'm using an Astrotrac, and while you mention you're not sure it'll work with it I don't see a problem as I'm using the TH3010 head which allows free movement of the RA axis--I assume I just have to take two photos 90-degrees apart to make this work, correct?

[I.C.D]

Well Richard I thought that I was the only one who had done that :lol::thumbsup::rofl: Ian .C

[themos]

Hello, I sent you a message, did you get it?

Themos

[lfc_star]

HI Themos,

I am quite new here. Are you still developing this software and is there any place I can download it to try ?

Thanks.
Nalin.

[bojan]

I just tried it, it doesn't work on my machine (w'7 Pro), it just shows the cmd window for a fraction od a second and then exits.

[Andy01]

I'm happy to give this a go, if it's mac compatible.

I can see the Celestial Pole region from where I usually set up.
I can attach a camera to my mount and bring image files to my laptop.
I can freely rotate the RA axis on my EQ6
I have internet connectivity

I don't have Python, is it also mac compatible?

Attached is a screenshot of my current PA, as seen through my mount, along with a phd graph.

Cheers

Andy

[themos]

Hello

I am not developing it anymore as it does all I intended for it! There are a couple of ideas for making it friendlier to people with optical trains more complex than camera+lens (automatic detection of horizon and image parity correction).

The source is available at https://github.com/ThemosTsikas/PhotoPolarAlign

It assumes you have a Python 2.7 installation on your machine ( I recommend the Anaconda distribution https://www.continuum.io/downloads) and it has been tested on Windows, Linux and Mac.

Themos Tsikas

[themos]

Hello,

If you have a python installation on your machine, you can open a command prompt window, go to (cd) to the directory where PPA is and run "c:\path\to\python PPA.py".

On my machine, for instabce, I can do

cd "C:\Program Files (x86)\PhotoPolarAlign"
c:\Users\Administrator\Anaconda\pyt hon.exe PPA.py

But I can also double click on PPA.py in the File Explorer or I can double click on a Shortcut I put on my desktop.

If you run it from a command line prompt, you will be able to see error messages to tell us what went wrong.

[themos]

Hello Andy,

I've had people with Macs use it. As I say, get Anaconda for Mac and then you can run all kinds of python programs, such as this one. It's also been used by Southern Hemisphere people.

Themos Tsikas

[bojan]

Result is attached...

[themos]

Hello,

I think this is a symptom of using python3 instead of python2.7. Can you check which version of python you are using?

Themos

[Visionoz]

Hi Themos

Can attest to the fact the your PPA works perfectly for me since when you posted here and I went to Stargazers to get some help from you initially about setting it up and all.

Agree that it does what you designed it for and helped me to polar align my EQ6 when out mobile and away from obsy. Using Win7 Pro 64bit & EOS 40D with a 100mm macro as the lens.

Cheers and once again thanks for the open-source PPA util mate.:thumbsup:

HTH
Bill

[bojan]

It is latest, 3.4.. I need it for some other apps.

What's next move?

[themos]

Hello,

I think that both versions can coexist, so download 2.7 Anaconda from https://www.continuum.io/downloads#_windows

Themos

[themos]

Hello bojan,

I just thought I would ask if you had any success with it yet?

Themos

[bojan]

Hi Themos,
No luck so far.. I couldn't make two versions of python coexist, so I gave up for now..
Could you try to adjust your code for the latest versions of python?

[OICURMT]

I assume you mean freeware and not public domain. You might consider using the GPL licensing to protect yourself from an intellectual property standpoint. If you make it public domain, anyone can modify it without recognizing the original works and sell it regardless.

I took a look at the source, while you have the authour listed, you really need to include a copyright notice and make yourself the copyright holder.

I suggest the following...

Created on Sun Oct 12 22:40:05 2014

Copyright 2016 - Themos Tsikas, Jack Richmond
All right reserved
or include the standard GPL license located at http://www.gnu.org/licenses/gpl-3.0.en.html

If you decide to GPL it, bear in mind that people can still modify it, but must include the original source and make reference to the original authours / copyright holders.

The other thing it does is protect you in the way of warranties. If your program screws something up, you will not be held liable. The bottom of the page explains how to include the license into your software.

Look for "How to Apply These Terms to Your New Programs"


OIC!
Standard Disclaimer. I am not a lawyer, my advice is worth exact what you are paying me... in this case, zero. :P

[themos]

That's a shame, I thought it was possible. I am not sure when I might get round to investing the time necessary to port it to python3.

Themos

[themos]

I actually meant public domain! please make money from this, if you can, that will be a hoot.

Themos

[OICURMT]

WOW.. you don't see that often

[torsion]

Hi Themos,

Just wanted to say thanks. After installing Python via Anaconda all is working great (tried it within cygwin with no luck).

I have an iOptron iEQ45 mount with the usual ALT and AZ adjustment screws and nuts. I wrote a little python script to tell me how many turns to each adjustment screw I 'should' apply to optimise the polar alignment (this requires a few physical parameters from the mount, such a thread pitch, radius of actuation etc).

After PPA gives the 'Move' details, I put these into my little script and it tells me which knob to turn and by how much. I thought I added (or incorporate) it to your program, but your python skills are well above mine so I will leave that for another day.

Attached is a screenshot after three iterations with the suggested turns (did some PA improvements after a few days).

My polar alignment routine is greatly simplified and improved, thanks heaps.

cheers,
Bram

[themos]

Hello Bram, that's good to know. I am a bit worried by the quality of the astrometry solve though, the stars are not centred in their circles, or at least they don't seem to be. Can you share your actual images, please?

Themos

[torsion]

Hi Themos,

Yeah, I noticed that but didn't think much of it. I upload the images onto the nova server and await the response.

Attach are my last three 'improvements', they are on flikr

1) https://flic.kr/p/QYwo2r
2) https://flic.kr/p/R6GvxG
3) https://flic.kr/p/R6Gv7w

Thanks for checking this out.

cheers,
Bram

[themos]

Hello, I submitted one of your images to nova.astrometry.net and I see this:

Center (RA, Dec):(50.025, -85.317)
Center (RA, hms):03h 20m 06.087s
Center (Dec, dms):-85° 19' 02.717"
Size:26.7 x 17.8 deg
Radius:16.071 deg
Pixel scale:47 arcsec/pixel

I think that's too wide an image, frankly. The more sky you include, the harder it is to match stars as optical distortion kicks in. I would try using a different optical system with less than half that scale, that is, under 20 arcsec/pixel and I would be more comfortable with something under 10. About 3 fingers of sky should be enough.

Themos

[torsion]

Thanks. Hmm, why didn't I think of that.

This is my (our) only lens for our canon, which piggybacks on my telescope to do wide field shots. Always wanted to get a longer focal length less but haven't manage to justify:)

Maybe I will see to do a drift align and see how far off the alignment is, given the wide field of view.

cheers,
Bram

[torsion]

I had the photos linked to the astrometry.net group on flikr and just received their results, and they say 27 arcsec/pixel.

However, if you do the math (or use an online FOV calculator), they give ~17 arcsec/pixel for a Canon EOS T2i with a 50mm lens. So I am confused.

[Amaranthus]

Hi Themos. This looks like a neat utility. But I wonder if you had a list of the Python libraries it requires? I have worked out the following, by debugging, but I might not have caught them all: numpy, scipy, astropy, ujson, pillow (PIL), image

[Wavytone]

Themis, your method for doing this is kind of old-tech.

I’m able to connect my iPhone via wifi directly to my camera, which provides full remote control including live view, taking the shot and transferring the image to the iPhone. There is also an api for iOS developers.

Seems to me it would be far simpler to write an iOS app that does it all in the phone or iPad.

Maybe one day someone will.

[themos]

It took a few years but it seems that someone has: the ASIAir Pro device seems to incorporate many bits of software we all use, including the functionality of PPA.