Bug 414128 - Solver failed with ASTAP
Summary: Solver failed with ASTAP
Status: RESOLVED UPSTREAM
Alias: None
Product: kstars
Classification: Applications
Component: general (show other bugs)
Version: 3.3.7
Platform: Other macOS
: NOR normal
Target Milestone: ---
Assignee: Jasem Mutlaq
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2019-11-14 10:35 UTC by Jean-Claude
Modified: 2019-11-25 13:10 UTC (History)
1 user (show)

See Also:
Latest Commit:
Version Fixed In:
Sentry Crash Report:


Attachments
solution.ini (591 bytes, text/plain)
2019-11-14 13:56 UTC, Jean-Claude
Details
Failed solution.ini (227 bytes, text/plain)
2019-11-14 14:03 UTC, Jean-Claude
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Jean-Claude 2019-11-14 10:35:09 UTC
SUMMARY
I didn't get any success using the ASTAP solver included in the last version of KStars.
The same images are solved with the Astro.net solver.
The same images are also solved directly with ASTAP (no KStars).
I using the last version of ASTAP.

STEPS TO REPRODUCE
1. 
2. 
3. 

OBSERVED RESULT


EXPECTED RESULT


SOFTWARE/OS VERSIONS
Windows: 
macOS: macOS 10.14
Linux/KDE Plasma: 
(available in About System)
KDE Plasma Version: 
KDE Frameworks Version: 
Qt Version: 

ADDITIONAL INFORMATION
Comment 1 han.k 2019-11-14 12:02:41 UTC
Binning of the exposed image provided to the solver could be a problem. Set the value behind bin: ... lower. The resulting image should be 1000 pixels high or more.

The downsampling (astap settings) is set best at value 0. (auto selection)

Han
Comment 2 Jean-Claude 2019-11-14 12:31:03 UTC
Hello Han,
Binning = 1x1, Downsampling = 0
However I got no improvement using the Simulator.
While the Astro.net solver worked flawlessly , the ASTAP solver failed each time when used in KStars.
Comment 3 han.k 2019-11-14 12:45:51 UTC
If the images are solved in ASTAP directly, it must be something with the EKOS settings. What are the other ASTAP settings in EKOS?

Could you upload an image to nova.astrometry.net and give me the link, so I can inspect the image?

Han
Comment 4 Jean-Claude 2019-11-14 13:05:44 UTC
Here are the EKOS options:
- In the Solver panel : -r 30 -speed auto
- In the Configure / ASTAP panel : 
Full path to the executable = /Applications/ASTAP.app/Contents/MacOS/astap
Search Radius = 30 checked,
Down Sample = 0 unchecked,
Update FITS unchecked,
Force Large Search Window unchecked
Comment 5 han.k 2019-11-14 13:12:39 UTC
That looks all good. I like to investigate the contains of the .ini text file written at the same location as the image to solve. This file contains the command line used.

I have written ASTAP, but are unfamiliar with EKOS (and Mac). Any idea where the image file and .ini are located and if so can you send them to me?
Comment 6 Jean-Claude 2019-11-14 13:21:07 UTC
Han, there is no .ini file created after a Solve, if the solve has been issued from EKOS or directly from the app ASTAP
I have uploaded a FITS file to http://dl.free.fr/fy08m4NIx
Regards,
Comment 7 han.k 2019-11-14 13:38:08 UTC
Jean-Claude,

That image is with 30 seconds exposure is very easy to solve. 

Looking to Ekos in my Linux virtual machine, EKOS store the image to solve at /tmp/solution

The solution or failure is reported in file /tmp/solution.ini 

Could you search for file solution.ini and report the text contains. It contains someting like:

PLTSOLVD=T
CRPIX1= 1.1645000000000000E+003
CRPIX2= 8.8050000000000000E+002
CRVAL1= 1.0672122122068352E+001
CRVAL2= 4.1224072440508330E+001
CDELT1=-7.4834768164008759E-004
CDELT2= 7.4831862727806173E-004
CROTA1=-1.7903448454242167E+002
CROTA2=-1.7903535373932309E+002
CD1_1= 7.4824162104277934E-004
CD1_2= 1.2609637447165383E-005
CD2_1= 1.2598775947195744E-005
CD2_2=-7.4821237959203386E-004
CMDLINE=/usr/local/bin/astap -r 30 -f "/home/h/astap.fpc/M31_50s_2x2_20171025_230034  0xFD  0xF  90xD  0x0R  0x0G  0x0B  0x0RGB  5x50L  _stacked normal.fit" -o /tmp/solution


The last line tells me the command used and the image to solve.  Either the command is wrong or the image.
Comment 8 han.k 2019-11-14 13:51:53 UTC
Correction EKOS uses also other names for the FITS file and .ini file. So you can only look for the extension .ini of .fits

I will try to install EKOS on my macOS simulator.
Comment 9 Jean-Claude 2019-11-14 13:53:13 UTC
If needed, you could try a TeamViewer session on my machine.
As you wish.
Comment 10 Jean-Claude 2019-11-14 13:56:02 UTC
Created attachment 123924 [details]
solution.ini
Comment 11 Jean-Claude 2019-11-14 13:56:21 UTC
Here is the solution.ini file
Comment 12 Jean-Claude 2019-11-14 14:03:01 UTC
Created attachment 123925 [details]
Failed solution.ini

Here is a solution.ini file after a solver failure
Comment 13 han.k 2019-11-14 14:04:24 UTC
The .ini file is from an succesfull ASTAP solve. ???

At the moment I'm also struggeling with the Linux version of Ekos to get the simulation working. I could try later to log in your machine to have look. How much more time do you have now?
Comment 14 Jean-Claude 2019-11-14 14:08:27 UTC
We can connect for the next two hours.
The first solution.ini file I sent was for a successful solver, the second was for a failed solver
Comment 15 Jean-Claude 2019-11-14 14:11:49 UTC
Han, somewhere in your code you may initiate an opening of the app ASTAP.
May be this should not be done because the app ASTAP opens the data base needed by EKOS . This causes to a failure.
Comment 16 han.k 2019-11-14 14:54:52 UTC
Hello Jean-Claude,

Your last remarks I don't understand. Can you explain it in more detail.


I have now created a simulation where ASTAP can solve the image loaded in the viewer but not the EKOS commandline. I'm now investigating why this is happening. 

Han
Comment 17 han.k 2019-11-14 15:03:31 UTC
Something goes wrong with the commandline. This needs further investigationn:


h@h-virtual-machine:~$ /home/h/astap.fpc/astap -r 10 -speed auto -f /tmp/fitsO34267.fits -o /tmp/solution

(astap:34831): GLib-GIO-CRITICAL **: 16:01:15.695: g_dbus_proxy_new: assertion 'G_IS_DBUS_CONNECTION (connection)' failed
h@h-virtual-machine:~$ /home/h/astap.fpc/astap -r 10 -speed auto -f /tmp/fitsO34267.fits -o /tmp/solution
Comment 18 Jean-Claude 2019-11-14 15:10:24 UTC
As I understand it when I use the ASTAP solver WITHIN EKOS, this should not call the independent app ASTAP.
This is however the case and the app ASTAP opens the G17 Star Catalog ... which prevents EKOS to access it.
Hence we get a solver failure.
Comment 19 Jean-Claude 2019-11-14 15:11:47 UTC
Han, let's make a TeamViewer session and you'll understand immediately
Comment 20 han.k 2019-11-14 15:23:04 UTC
no problem. What is de link
Comment 21 han.k 2019-11-14 15:35:38 UTC
You can send me a email with details to:
http://www.hnsky.org/hanemail.gif
Comment 22 Jean-Claude 2019-11-14 15:39:36 UTC
I sent you the data to han.k@hnsky.org
Comment 23 Jean-Claude 2019-11-14 15:40:15 UTC
My phone is +33 476 456 226
Comment 24 han.k 2019-11-14 16:48:05 UTC
Jasem,

We have tested again on Jean-Claude Mac. ASTAP solving works on real images, not on his simulated images. 

The images size of Jean-Claude in degrees are pretty small about 0.3 degrees. The simulated stars are not recognized by ASTAP. Maybe they are too big to be recognized as star. 

Solving an simulated image image of 1,27 by 1.02 degrees works here. 

Is it possible to reduce the star size in the simulation?

Han
Comment 25 Jasem Mutlaq 2019-11-25 10:40:29 UTC
The simulation driver is in INDI, not KStars. With v3.3.8 of KStars on Mac, does it work with ASTAP under Simulation?
Comment 26 han.k 2019-11-25 12:14:17 UTC
The original problem is solved by update for Indi, id 414155. The fact that ASTAP has some problems with solving a simulated image is just an observation. (ASTAP works better when intensity of stars is displayed)

I would suggest to close this bug report.