Plate solving and goto with AstroTortilla

[russellhq]

Downloading now

I take it that if you image with more than one scope you grab the index files in accordance with 20% the narrowest field and then the mex field? That's what its seems to imply and what i am doing.... So 3Gig of download. Wonder what speed this will run on my imaging laptop - a Core2 Duo with 4 Gig of RAM? Does this application swap to disk a lot? Could benefit from a SSD upgrade if that were the case.....

EDIT: The download site is very slow.... I will just leave it going all day and overnight if needs be

I posted earlier about the performance I experienced with AT:

After the images was captured, the different steps in the process were broken down roughly as follows:

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

As you see, the actual solve time where it searches the indexes only took 4s.

[Sp@ce_d]

Downloading now

I take it that if you image with more than one scope you grab the index files in accordance with 20% the narrowest field and then the mex field? That's what its seems to imply and what i am doing.... So 3Gig of download. Wonder what speed this will run on my imaging laptop - a Core2 Duo with 4 Gig of RAM? Does this application swap to disk a lot? Could benefit from a SSD upgrade if that were the case.....

EDIT: The download site is very slow.... I will just leave it going all day and overnight if needs be

I run it on a netbook.. runs ok.. seconds rather than minutes. That was before I upped it to 2GB & swapped in an SSD

[kirkster501]

hmmm, my download keeps dropping out when connecting to the "international server". I am on line all day with a 100Meg connection and all my real time work apps and my IP phone and Telepresense are working fine so its soemthign to do with the AT server's end. I may have to try later and try the Finland server.

Is there any way to download this onto one computer and move all the files to another anyone know please?

[Realt]

Is there any way to download this onto one computer and move all the files to another anyone know please?

The bulk of the download is the indices. You can download these seperately on any computer and transfer them into the correct folder on the machine you are going to use.

As others have saidAT is not particularly resource hungry and runs fine with all my other software on an xp netbook with 2gig ram.

Any slowness will be a parameter setting issue.

You do know that now the clouds will set in for the next month!

Clear skies here in Ireland but I've tweaked my back so no playing outside for me for a few days . Rats.

[Realt]

https://docs.google....VmhaQ1Z4UHlOWlk

image scale is between 5 and 6 arcseconds per pixel.

@kirkser501

If you just download the indices needed to solve this you can be checking your set up while you wait on the rest.

Its good to use this jpeg as you know it will solve if all is ok and you can compare your logfile with the ones elsewhere in this thread-it helped me a lot.

[kirkster501]

I am having issues with the download server resetting. Tried on different computers and on other broadband connections and just keeps crashing out ... Maybe try tomorrow.

@Realt. Which of the indices do I need to solve that picture please?

[Realt]

Using the parameters shown on the attached screengrab you can see from the log that it solved from 4014,

parameters.zip

Log.zip

Other log in thread (look near post from themos I quoted) solved same jpeg from 4010 but sigma different I'd guess.

However these are the (MUCH) smaller index files so if you went a few either side of 4014 you should be good for a bit of experimenting I'd say.

best of luck.

[Realt]

http://broiler.astrometry.net/~dstn/4000/

link direct to indices.

might be easier to get them from here and put them in the correct folder in cygwin-or maybe not.

[kirkster501]

Doh.... giving up until tomorrow. Maybe the server will have settled down by then.

[Realt]

And make sure you download the full size jpeg.

[kirkster501]

Thanks for the direct download link. Just gonna grab as many as I can overnight. Shame they don't run a ftp server on that site.

[kirkster501]

Well, the bz2 files (teh format the indices download from the astronometry website) uncompress with a simple double click on my mac to .fits files. So I would assume I just bung all the fits files into the required directory and bobs your uncle.... I will play and look later. I am supposed to be working from home today...!

[Realt]

Well, the bz2 files (teh format the indices download from the astronometry website) uncompress with a simple double click on my mac to .fits files. So I would assume I just bung all the fits files into the required directory and bobs your uncle.... I will play and look later. I am supposed to be working from home today...!

That's correct.

Default location is

C:\cygwin\usr\share\astrometry\data

but as somebody pointed out earlier in the thread you could put them elsewhere and point AT at them if you needed to be complicated.

Work is over rated -do some science instead.

[kirkster501]

@Realt / guys I used exactly the same settings as you with that image you link to and it worked in 16 Seconds !! Way to go!

So, confidently, I try this on one of my own images, and that's where it goes wrong....!

I am using some of my old lights (Rosette Nebula and quite heavy LP) to check this from my ED80 with FR with DSLR 450d - a combination that has a FoV of 180 arc mins by 120 - give or take a few minutes.

So in the AT main window, I have changed the Scale Units to "arcmins" and "scale minimum" to 24 (i.e. 20% of the minimum scale of 120 in minutes) and "scale maximum" to 180 - i.e. the maximum in minutes. Does that sound logical? Of course, with another scope it would be a different set of values.

Either way, it does not solve with indices 4003 through to 4012.

I think I am nearly there - its something to do with these settings that are stopping it working.... Of course, under the skies is the true test but in our weather when will that be ?

Below is my AT window. I am not sure what to adjust TBH... I have tried with a few pictures and it does not work with any. If there are any more example pictures folks have with different scales for me to try I will appreciate if you can share to check my set up out.

Thanks, Steve

[Realt]

I would set minimum scale closer -say 170 arcmins.

set downscaling to 0

gradually increase sigma until it extracts about 100 objects

Thats just where I would start.

If you post a full logfile somebody will be able to give you some better pointers.

Sometimes it just won't solve so don't get too disheartened.

Somebody did post another image earlier in thread with a logfile of a successful solve.Look back through thread.

If you dump your image somewhere I will download and have a play-I'm far from an expert on AT yet though.

[kirkster501]

Appreciate soemone quickly checking out my log please? Mayb3 its just my image, here (4.7Meg) if anyone would like to look? Image scale is 2.65 x 1.77 degrees.

https://dl.dropbox.c...89/rossette.jpg

2013-02-08 21:02:33,750 - astrotortilla - INFO - Connecting to camera...

2013-02-08 21:02:33,750 - astrotortilla - INFO - Exposing: 5.00 seconds

2013-02-08 21:02:37,842 - astrotortilla - INFO - Reading image from camera

2013-02-08 21:02:37,842 - astrotortilla - INFO - Solving...

2013-02-08 21:02:38,765 - astrotortilla - INFO - Reading input file 1 of 1: "/cygdrive/c/Documents and Settings/User/Desktop/rossette.jpg"...

2013-02-08 21:02:39,483 - astrotortilla - INFO - jpegtopnm: WRITING PPM FILE

2013-02-08 21:02:41,608 - astrotortilla - INFO - Read file stdin: 4272 x 2848 pixels x 1 color(s); maxval 255

2013-02-08 21:02:41,703 - astrotortilla - INFO - Using 8-bit output

2013-02-08 21:02:47,358 - astrotortilla - INFO - Extracting sources...

2013-02-08 21:02:51,296 - astrotortilla - INFO - simplexy: found 4543 sources.

2013-02-08 21:02:53,312 - astrotortilla - INFO - Solving...

2013-02-08 21:02:54,921 - astrotortilla - INFO - Reading file "/cygdrive/c/docume~1/user/locals~1/temp/solvermuljr4/4/rossette.axy"...

2013-02-08 21:02:55,030 - astrotortilla - INFO - Field 1 did not solve (index index-4013.fits, field objects 1-10).

2013-02-08 21:02:55,733 - astrotortilla - INFO - Field 1 did not solve (index index-4012.fits, field objects 1-10).

2013-02-08 21:03:03,358 - astrotortilla - INFO - Field 1 did not solve (index index-4011.fits, field objects 1-10).

2013-02-08 21:03:03,467 - astrotortilla - INFO - Field 1 did not solve (index index-4010.fits, field objects 1-10).

2013-02-08 21:03:03,562 - astrotortilla - INFO - Field 1 did not solve (index index-4009.fits, field objects 1-10).

2013-02-08 21:03:03,671 - astrotortilla - INFO - Field 1 did not solve (index index-4008.fits, field objects 1-10).

2013-02-08 21:03:03,780 - astrotortilla - INFO - Field 1 did not solve (index index-4007-11.fits, field objects 1-10).

2013-02-08 21:03:03,905 - astrotortilla - INFO - Field 1 did not solve (index index-4007-10.fits, field objects 1-10).

2013-02-08 21:03:04,015 - astrotortilla - INFO - Field 1 did not solve (index index-4007-09.fits, field objects 1-10).

2013-02-08 21:03:04,125 - astrotortilla - INFO - Field 1 did not solve (index index-4007-08.fits, field objects 1-10).

2013-02-08 21:03:04,233 - astrotortilla - INFO - Field 1 did not solve (index index-4007-07.fits, field objects 1-10).

2013-02-08 21:03:04,328 - astrotortilla - INFO - Field 1 did not solve (index index-4007-06.fits, field objects 1-10).

2013-02-08 21:03:04,437 - astrotortilla - INFO - Field 1 did not solve (index index-4007-05.fits, field objects 1-10).

2013-02-08 21:03:04,530 - astrotortilla - INFO - Field 1 did not solve (index index-4007-04.fits, field objects 1-10).

<LOADS MORE - SNIPPED all teh way down to 4003 index>

2013-02-08 21:03:06,015 - astrotortilla - INFO - No solution in 28.2s

[russellhq]

OK, I tried your photo and it seemed to solve no problem.

I used arcsec/pixel as the scale units, so used this site to work it out for your setup.

http://www.wilmslowa...re/formulae.htm

I assumed 510mm for your FL (based on 80ED with 0.85x FR) and a pixel size of 5.2um for your 450D. This gave an answer of 2.1, so I set the min as 2.0 and max as 2.2.

Below is a copy of the cfg file for info.

[Telescope-ASCOMTelescope]
slewsettletime = 0.0
syncaccuracy = 1.0
lastselection = EQMOD_SIM.Telescope
syncmaxwait = 2.0
propertyagelimit = 1.0
[Camera-ASCOMCamera]
binning = 1
lastcamera =
[Camera-APTCamera]
hostname = localhost
port = 21701
binning = 1
[Solver-AstrometryNetSolver]
shell = C:\cygwin\bin\bash.exe --login -c "%%s"
shell_hide = True
downscale = 4
scale_max = 2.2
scale_xrefine = 0
configfile = /etc/astrometry/backend.cfg
scale_low = 2.0
xtra = --sigma 1 --no-plots --overwrite -N none
searchradius = 180
year_epoch = JNOW
scale_units = arcsecperpix
[Session]
logwindoww = 698
syncmode = 0
guiy = 105
solver = AstrometryNetSolver
guiw = 650
iterlimitcount = 5
guix = 717
logwindowx = 179
logwindowy = 85
camera = APTCamera
guih = 511
iterlimitarcmin = 1.0
logwindowh = 292
exposure = 5.0
[Camera-FileOpenCamera]
directory = C:\APT_Images\Camera_1\2012-10-26
binning = 1
[Bookmarks]
count = 0
[AstroTortilla]
work_directory = C:\AstroTortilla\Temp
settings_path = C:\Users\Russell\AppData\Local\astrotortilla.sf.net\AstroTortilla
log_file = Solvelogfile.log
log_level = INFO
last_gotoimage = C:\Users\Russell\Downloads\rossette.jpg

[kirkster501]

Thanks Russell.

So in the AT app window you changed to arc seconds in there? I have not seen this cfg file approach as yet - RTFM time I think. I see your scale max is 2.2 but why? if its secs you are using then 2.7 degrees (the image scale of that light sub) - 2.7x60x60 = 9720? Why not just use degrees?

Seems a little bit of a dark art what values to use and i am trying to see some pattern.

Appreciate the ongoing steers here guys.

Steve

[russellhq]

You can save your current settings and load them later, it's quite handy for different imaging set-ups.

I set the arcsecperpixel in the AT window, see below;

I used 2.2 because it was 0.1 more than the calculated scale (2.1), I was trying to keep the range tight to keep things moving quickly. If it hadn't worked I probably would have tried 2.3 and so on until it worked.

I used seconds as it was quite easy to calculate using the link I posted. How did you calculate the image scale?

[kirkster501]

Ahhh, right. Am I going nuts here? When you refer to scale you mean the pixel size? I thought scale was the actual degree and minutes the image represented in the sky - the scale represents the angular size the image hints to AT which index to pick to try and get a match? But in fact the index is picked my the pixel scale? I shall look into that if so.

[Realt]

Ahhh, right. Am I going nuts here? When you refer to scale you mean the pixel size? I thought scale was the actual degree and minutes the image represented in the sky - the scale represents the angular size the image hints to AT which index to pick to try and get a match? But in fact the index is picked my the pixel scale? I shall look into that if so.

As I unerstand it arcseconds per pixel and degreewidth are just two ways of saying the same thing-you can use the units that suit you.

i.e width of the image in degrees divided by no of pixels on sensor equals arc/sec per pixel.

Can somebody correct this if i am wrong.

Setting the scale max and min in whatever units you choose is just narrowing the boundaries of the search for a solve.

@russellhq.

I can't get this image to solve.what sigma did you use and did you downscale by 4??

[russellhq]

AT has three units of scale you can use:

scale_units degwidth, arcminwidth or arcsecperpix

I used sigma 1 and downscaled by 4.

[Realt]

Right. Solved it.

All parameters shown on attached screen grab.

As you see I used degwidth.

I am still trying to figure out the downscaling settings. Must rtfm again.

solved.zip

[Realt]

AT has three units of scale you can use:

scale_units degwidth, arcminwidth or arcsecperpix

I understand.

I was just pointing out to kirkster501 that these are the same thing effectively as he didn't seem to be grasping it.

[kirkster501]

Thanks it solved ! 31 secs.

I need to get my head around the values to input I think and read the documentation a little more. At least I am set up and configured correctly!

What is odd though is AT changes the min and max values after a solve. And then when i run the same image again it does not solve with those values.....

Thanks Russ and Realt !