Plate solving and goto with AstroTortilla

[Earl]

i starting from scratch again and have deleted everything. also downloaded the version 3 which installed without error and did the cygwin stuff ok this time

[Earl]

Progress im now getting the field 1 did not solve message so its at least working to a degree

[Earl]

Do I have success?

Im running via Ascom Simulator and CdC for the scope and use the tools goto image option.

The plate solves and the scope slews to the correct image.

I guess the second part of this is to get to to then take a sub and get better alignment?

[Earl]

I have part one sorted I think.

I just instralled it on the observatory PC and performed a image goto from my last M45 data and it solved and slewed, with the cloud I am unable to go beyond that but so far so good

[themos]

yes, that's good!

[Earl]

During the installation did you select the second download site?

That part thew me until I read the section again on exactly what to do, I was typing in the example, not ctrl clicking another server.

Its not that complicated now I have worked it out, its just a little more work than normal.

[RogerTheDodger]

Excellent, glad you have it working.

[zimzimma]

Hi all,

Ive downloaded this software namely for the Polar Alignment aspect. I have a couple of really noob questions but never the less ill ask them at risk of sounding dumb!

1)Will this work with my SPC900 webcam (I do use maximdl). If so which indexes would i require when on 200pds?

2)I understand if you show this software an image it will solve it and move the scope to that part of the sky, but im guessing without using the polar align function will it not literally track all over the place?

3)What are the advantages over this(other than time syncing etc)over doing a 3 point align/sync?

Thankyou again people.

Rt

[themos]

It should work with a webcam, I think I did some time ago. You get fits files from the webcam? if you want only some index files, you need to know the field of view of the camera. Your PDS has a 1000mm focal length which should give 9.2 x 12.3 arcminutes on the webcam. This is very small so you will need all the index files, essentially.

2) Once it solves an image, it can Sync the scope to the solved coordinates so the mount control knows where it is. This should help find specific targets nearby. In order to track reliably, the mount needs to be polar aligned first, anyway.

3) I don't know, I hardly ever do a multi point align!

[RogerTheDodger]

From a rough polar park position, no need to do a multi point align at all... just do first goto, snap image, analyse and re-slew, snap image to check. You are now on target.

AT has nothing to do with tracking, you need to do that with polar alignment as themos said.

[Sp@ce_d]

I haven't tried the Polar Alignment Tool or the Drift Shot Tool yet (my mounts on a pier in an obsy) but I should imagine they would be pretty accurate going by the results so far. The techniques are described at the end of the user guide.

As previously said, for imaging most people are only interested in alignment at the area of interest as they'll more often than not stay on that target all night. AT makes this a snap, it really is so cool.

[russellhq]

Started using this last night and worked quite well up to a point. I found that sometimes after an image had been solved the program hung trying to sync the new position and this caused all sorts of issues with EQMOD and trying to control the telescope. Has anyone else come across this problem?

[Sp@ce_d]

yes.. this has started with me the last couple of sessions. I've had to close CDC, Neb & AT pull the usb from the EQDir, reconnect & start them up again. Something is locking up sharing the bus I think. I never start PHD until finished solving as that doesn't like it. As long as I don't power off the mount I can pick up again quite quick but it does seem to have become fussier and I've made no recent changes to the setup either.

[Sp@ce_d]

Actually, I wonder if it's the camera interface as AT thinks it's waiting for a second frame to solve when it locks up..

[russellhq]

I'm using it with Stellarium Scope and APT. It does seem to hang when it's solving a second image. I had to restart the laptop to get it all going again, though didn't try unplugging the USB...

[russellhq]

I'm trying to get the annotated files after processing but having no luck. I dropped the --no-plots from the custom options section and the log window does show it's created the plot and tells me what it's found but I can't find the files?!

Anyone know how to get them?

[russellhq]

I've been playing about with AT, trying to get it to solve faster but noticed from the log the slowest part is actually opening the file and extracting the sources.

For example, it usually takes 20-25 seconds to solve an image, and from the log, that's broken down as:

4.5s to read and convert to pnm

11s to "Using 8-bit output" - I've no idea what that means or does...

3.5s to downsample by 4 (I found 4 was quicker than 2 or 1)

4s to solve

I think if the 8-bit output process could be sped up, then I could shave a good chunk of time off the overall solving process. Anyone else experiencing similar breakdown of durations?

Also, for solving, I used arcseconds/pixel to determine scale using the formula: [(pixel size in microns)/(focal length in mm)]*206.3 and set the min and max as 5% either side. I also left the search radius at 180 as narrowing it down to 5 made no difference to the solve times.

[russellhq]

Also, forgot to say I only have one set of indices in the data folder, 4006. I compared the solve times of the old and new index format (file names 206 and 4006) and found little difference in speed so I stuck with the new index

[russellhq]

Had some free time this morning so been playing about more with AT.

First I loaded up cygwin, typed "solve-field" and this gave me all the switches for using in the custom options box in AT. I then found out "--sigma" is used to set the background noise level in the image. Basically the higher the number the less stars will be identified. I found that setting it to 1 gave the fastest solve times.

I then found the option "--plot-scale", this will scale down your annotated image by the given factor. I used 0.125 and this greatly improved solve times (from 100s to 30s) when you want an annotated image.

Next up was the default "-N none" switch. This stops a new FITS file being created. Although I didn't experience any difference when it was there or not.

I also found that I had to add a line to the astrotortilla.cfg file (located below) to get the annotated files. The line was "work_directory = C:\AstroTortilla\Temp" under the[AstroTortilla] section

C:\Users\<Your Username>\AppData\Local\astrotortilla.sf.net\AstroTortilla\astrotortilla.cfg

The annotated file could the be found in the Temp folder added as the work directory.

I also started getting an error in the log about files already existing, so I added the switch "--overwrite" and this seemed to clear things up.

So, my custom settings are now:

For an annotated plot

"--sigma 1 --plot-scale 0.125 --overwrite -N none"

Without plot

"--sigma 1 --no-plots --overwrite -N none"

I wrote this mainly for my own benefit/reference but hopefully it will be of use to others.

[Black Knight]

They sound useful, thanks

[dark star]

I am getting some strange results when trying to centre objects using Astrotortilla. a couple of time it has moved objects further and further away, until they are not in the field of view at all!

I have entered Epoch J2000 in Astrotortilla, does anyone know if the latest Synscan Beta is JNOW, in case this is causing it?

David

[IanL]

I started trying to get set up with AT last night. I have the software installed, but now need to grab appropriate index files, which is where I ran in to some issues:

- The instructions provided on the sourceforge installation/user manual are out of date as far as I can see. You used to have to email someone to get a login to download the old '200' series index files, but the link to the instructions for emailing is now dead.

- Instead they now provide unrestricted access to the new '4000' series index files directly. Go here and follow the instructions: http://astrometry.ne...ing-index-files

- When you go to the index file download link, you will see there are a *lot* of index files that you can download. Don't download them all 'just in case' or your ISP will imprison you until you sell your children to pay for the excess bandwidth use. It may also slow down plate solving to a crawl if you have too many files of the wrong scale and you don't use the scale parameters in AT to give it a hint. You need to download index files that cover the appropriate range of scales for your camera/imaging rig combination.

- Firstly they say that you should aim to have index files that cover a range of 10% to 100% of your image scale. How do you figure that out? Well you can do the maths if you are inclined, or you can do it the lazy way.

- Get a full-frame image from your imaging rig with a reasonable number of stars in it. Do some basic preprocessing and a histogram stretch so you have nice clean stars and as little noise as possible. Don't do any stacking of images (which can expand or shrink the image a bit if you have drift between frames) and do not crop or adjust the image scale in any way. Save it as a GIF, JPG, PNG or FITS file as you see fit.

- Now upload the image to http://nova.astrometry.net/upload

- This will do a plate solve on the image for you (usually takes about ten minutes to complete). The page will auto-reload every ten seconds until it is ready, or you can manually refresh if it doesn't. Once ready, go to the result page, e.g.:

- Look at the "Radius" figure (in degrees); this is the distance between the centre of the image and the corners, i.e. tha radius of a circle that would pass through all four corners of the image. In this case the Radius is 1.478 degrees, so the largest feature scale we might want to solve for is twice the radius; 2 * 1.478 = 2.956 degrees.

- The index scales are given in arcminutes , so multiply by 60, giving 60* 2.956 = 177.36 arcminutes. We want indices between 10% and 100% of that scale, which means we require 17.736 to 177.36 arcminutes. Consulting the table at http://astrometry.ne...ing-index-files that tells us we need index sets:

index-4013.fits.bz2 : 170 - 240 arcminutes

index-4012.fits.bz2 : 120 - 170 arcminutes

index-4011.fits.bz2 : 85 - 120 arcminutes

index-4010.fits.bz2 : 60 - 85 arcminutes

index-4009.fits.bz2 : 42 - 60 arcminutes

index-4008.fits.bz2 : 30 - 42 arcminutes

index-4007-*.fits.bz2 : 22 - 30 arcminutes

index-4006-*.fits.bz2 : 16 - 22 arcminutes

Now your mileage may vary here:

Firstly you may not need all of the suggested image scales. If bandwidth is a problem to download, then I think I'd start with an index set in the middle of the suggested image scale range and try some tests (remember you can use AT's file loading camera to select old images from your hard disk, you don't have to test this bit under the stars).

Secondly, the online solver we used above reported that it actually solved my image using indices with a scale of 11-15 arcminutes, which is smaller than the suggested scales needed. The test image was highly processed however, and has lots of easily identifiable small stars. A single sub captured direct from your imaging software with no processing will likely need a larger scale image than the minimum suggested, hence my thought to start with mid-scale indices first.

- Once you have identified the appropriate image scales, you can further reduce your download mountain by not including index files for areas of the sky that are not visible from your location:

- The large scale index files (series numbers 08 through 19) are a single file covering the whole sky, so just download these if you need them.

- The medium scale index files (series numbers 05 through 07) are split in to 12 'tiles' covering different areas of the sky. If you look in the file 'hp.png' in the index downloads page, you will see a map of the RA and DEC covered by each tile. If you live a long way north, you may not need tiles 08 through 11 (or 00 through 03 if you live a long way south), so you could omit any tiles ending in that number.

- The small scale index files (series numbers 00 through 04) are split in to 48 'tiles', and again if you look in the file 'hp2.png' you can see which parts of the sky they cover and omit the ones that you can't or don't image from your latitude.

The final point to note is that once you have downloaded the files, the instructions to "tar xjf *.bz2" did not appear to work on my CygWin installation. It just complained about then not being tar archives. What I found is that using "bunzip2 *.bz2" works just fine. Not sure why, but worth knowing!

[Sp@ce_d]

Nice write up. Thanks for doing this.. I haven't got around to trying out the new index files yet (if it ain't broke why fix it).

Now you've done this I think I'll give them a go & see if I can speed it up even more.

[Mik3]

I cant figure out what indices to download, I tried the steps above but cant get it to work can anyone help me?

I use a canon 1000 D and Sky Watcher 200P DS.

I used a fov calculator and apparently my fov is 1.27° x 0.85°

[themos]

I'd start with the indices

• index-210+.tar.bz2 60 - 2000 29
  
• index-209.tar.bz2 42 - 60 30
  
• index-208.tar.bz2 30 - 42 61
  
• index-207.tar.bz2 22 - 30 125
  
• index-206.tar.bz2 16 - 22 255
  
• index-205.tar.bz2 11 - 16 526
  
• index-204.tar.bz2 8 - 11 1,023
  
• index-203.tar.bz2 5.6 - 8.0 2,089