×

INDI Library v2.0.7 is Released (01 Apr 2024)

Bi-monthly release with minor bug fixes and improvements

Read More...

Alignment/Plate solve problems

  • Posts: 47
  • Thank you received: 0

Replied by Steen on topic Alignment/Plate solve problems

I was just taking a snapshot of the pole. Here are two examples where I try to solve against Pollux and Regulus. If I use the simulator everything is ok. If I use gphoto the alignment fails. I have enclosed tar-ball incl. fits of simulator and my canon used for solving. I can from these tell that I'm pointing at the right star and thus are within 30 degrees.

I was wondering if the faillure could have something to do with debayering?

/steen

PS I cant upload the pics due to size
Last edit: 9 years 1 month ago by Steen.
9 years 1 month ago #3629
Attachments:

Please Log in or Create an account to join the conversation.

  • Posts: 43
  • Thank you received: 1

Replied by Marko T. on topic Alignment/Plate solve problems

So shorten the story (in the right place :blush: ). My original problem was that the solver refused to start solving and gave a message "Illegal instruction (core dump)" before the extraction of sources starts. The same happened in command line and ekos.

This problem appeared after updating to the 0.52 version of astrometry.net from the repo. This most likely does not affect everyone, so if you have problems like in the original post, this could help.

The solution was to compile the astrometry.net package according to the instructions in the HOW TO thread. In addition uncomment the "export OPTIMIZE = no" line in the makefile.

Still very happy with kstars/indi/ekos even though there are some bumps on the road :woohoo:
9 years 1 month ago #3648

Please Log in or Create an account to join the conversation.

Replied by Jasem Mutlaq on topic Alignment/Plate solve problems

Btw, the package actually turns optimization off, but it seems this does not have any effect.
9 years 1 month ago #3651

Please Log in or Create an account to join the conversation.

  • Posts: 43
  • Thank you received: 1

Replied by Marko T. on topic Alignment/Plate solve problems

Last astrometry.net update "killed" solver for me again. Got it working again with recompiling, but I had to do "make clean" and then "sudo make install INSTALL_DIR=/usr FINAL_DIR=/usr ETC_INSTALL_DIR=/etc DATA_FINAL_DIR=/usr/share/astrometry DATA_INSTALL_DIR=/usr/share/astrometry SYSTEM_GSL=yes py extra". Just giving the "sudo make install" command didn't work after "git pull".
9 years 1 month ago #3661

Please Log in or Create an account to join the conversation.

  • Posts: 43
  • Thank you received: 1

Replied by Marko T. on topic Alignment/Plate solve problems

Still some woes (but getting there).

On the command line with the exact same parameters the solver does what it should, but running (load and slew) with the same parameters and image in ekos does not work.

Ekos dialog
2015-03-20T17:27:09 Solver failed. Try again.
2015-03-20T17:27:09 Reading input file 1 of 3: ""...
Base: "./.", basefile ".", basedir ".", suffix "(null)"
Checking if file "" ext 0 is xylist or image: image
  (not xyls because: Failed to open FITS table : Failed to open FITS file "")
Running: /usr/bin/image2pnm.py --no-fits2fits --fix-sdss --infile  --uncompressed-outfile /tmp/tmp.uncompressed.hbUNrW --outfile /tmp/tmp.ppm.zQJFg4 --ppm
2015-03-20T17:27:09 /usr/bin/solve-field --verbose --no-verify --no-plots --no-fits2fits --downsample 2 -u aw -r -L 74.2325 -H 123.07   -W /tmp/solution.wcs /home/mtoastro/Pictures/m51_1min_011.fit
The same parameters and image in command line:
Reading input file 1 of 1: "Pictures/m51_1min_011.fit"...
Base: "Pictures/m51_1min_011", basefile "m51_1min_011.fit", basedir "Pictures", suffix "fit"
Checking if file "Pictures/m51_1min_011.fit" ext 0 is xylist or image: image
  (not xyls because: FITS file does not have any extensions)
Running: /usr/bin/image2pnm.py --no-fits2fits --fix-sdss --infile Pictures/m51_1min_011.fit --uncompressed-outfile /tmp/tmp.uncompressed.1lpVeu --outfile /tmp/tmp.ppm.QCKQyE --ppm
  fits
Running: pnmfile /tmp/tmp.ppm.QCKQyE
Extracting sources...
Running image2xy: input=Pictures/m51_1min_011.fit, output=/tmp/tmp.xyls.vjqssd, ext=0
nhdus=1
Got naxis=2, na1=3906, na2=2604
Downsampling by 2...
simplexy: nx=1953, ny=1302
simplexy: dpsf=1.000000, plim=8.000000, dlim=1.000000, saddle=5.000000
simplexy: maxper=1000, maxnpeaks=10000, maxsize=2000, halfbox=100
simplexy: median smoothing...
simplexy: measuring image noise (sigma)...
Sampling sigma at 6370 points
Nsigma=0.7, s=12.5214
{removed bunch of lines}
Spent 38.2529 seconds on this field.
Field: Pictures/m51_1min_011.fit
Field center: (RA,Dec) = (202.4, 47.2) deg.
Field center: (RA H:M:S, Dec D:M:S) = (13:29:44.679, +47:11:51.843).
Field size: 1.9683 x 1.31185 degrees
Field rotation angle: up is 51.223 degrees E of N
Creating new FITS file "Pictures/m51_1min_011.new"...
Reading input file FITS headers...
Regular expression matched: "^END$", key "END".
Reading WCS file FITS headers...
Regular expression matched: "^SIMPLE$", key "SIMPLE".
Regular expression matched: "^BITPIX$", key "BITPIX".
Regular expression matched: "^NAXIS$", key "NAXIS".
Regular expression matched: "^EXTEND$", key "EXTEND".
Regular expression matched: "^END$", key "END".
Copying from offset 2880 to offset 40688640 (length 40685760) of the input file to the output.
Deleting temp file /tmp/tmp.ppm.QCKQyE

Running kstars in debug mode does not supply any additional info on the situation.
mtoastro@mtoastro-K53U:~$ uname -r
3.19.0-9-generic
mtoastro@mtoastro-K53U:~$ dpkg -l astrometry*
Desired=Unknown/Install/Remove/Purge/Hold
| Status=Not/Inst/Conf-files/Unpacked/halF-conf/Half-inst/trig-aWait/Trig-pend
|/ Err?=(none)/Reinst-required (Status,Err: uppercase=bad)
||/ Name           Version      Architecture Description
+++-==============-============-============-=================================
ii  astrometry-dat 0.45         all          Astrometry.net index files
ii  astrometry-dat 0.45         all          Astrometry.net index files
ii  astrometry-dat 0.45         all          Astrometry.net index files
ii  astrometry-dat 0.45         all          Astrometry.net index files
ii  astrometry-dat 0.45         all          Astrometry.net index files
ii  astrometry.net 0.52+r3790.2 amd64        Astrometry plate solver
mtoastro@mtoastro-K53U:~$ dpkg -l kstars*
Desired=Unknown/Install/Remove/Purge/Hold
| Status=Not/Inst/Conf-files/Unpacked/halF-conf/Half-inst/trig-aWait/Trig-pend
|/ Err?=(none)/Reinst-required (Status,Err: uppercase=bad)
||/ Name           Version      Architecture Description
+++-==============-============-============-=================================
un  kstars         <none>       <none>       (no description available)
ii  kstars-bleedin 5:15.04+r516 amd64        desktop planetarium for KDE
ii  kstars-bleedin 5:15.04+r516 all          data files for KStars desktop pla
ii  kstars-bleedin 5:15.04+r516 amd64        debug information for the desktop
un  kstars-data    <none>       <none>       (no description available)

Any ideas?
9 years 4 weeks ago #3671

Please Log in or Create an account to join the conversation.

Replied by Jasem Mutlaq on topic Alignment/Plate solve problems

email me your FITS file. mutlaqja AT ikarustech DOT com
9 years 4 weeks ago #3672

Please Log in or Create an account to join the conversation.

  • Posts: 47
  • Thank you received: 0

Replied by Steen on topic Alignment/Plate solve problems

Lastest build form the repos does also segfault for me.

I have tried to build astrometry.net again and found that it depends on netpbm and the version on my system (xubuntu 14.10 with daily builds) didn't have the right version.

Insted I compiled and build netpbm from SVN and build astrometry.net 0.50 from the astrometry site. Then I can sove again both the demos and with the indi_simulator_ccd. But again when I use canon 1100D and gphoto the solver fails.

Has someone a clean system working, where plate-solving is working in Ekos with gphoto. Then I would try to find the difference to my setup...

Please supply me with versions on your working setup (ubuntu, kstars, ekos, libindi, astrometry.net, netpbm)
9 years 4 weeks ago #3673

Please Log in or Create an account to join the conversation.

  • Posts: 43
  • Thank you received: 1

Replied by Marko T. on topic Alignment/Plate solve problems

I'd really like to know too. I was looking at logs when I'm running the solvers in command line and ekos. There seems to be a difference in the order things are performed.
command-line
Read input file 1 of 1
Base
Run image2pnm
Run image2xy
(Solving ontinues)

Ekos
Run image2pnm
Base
Read input file 1 of 3
(solver fails)

The order of the solving process is quite different. This might be due to different amount of verbosity you get in ekos and command prompt. Reading of input files in ekos apparently fails because the tmp -file it tries to create with image2pnm is empty. The tmp -files created by the command line solver are actual images.
Last edit: 9 years 4 weeks ago by Marko T..
9 years 4 weeks ago #3674

Please Log in or Create an account to join the conversation.

Replied by Jasem Mutlaq on topic Alignment/Plate solve problems

Please email me a screenshot of the alignment module plus your canon 1100D image. mutlaqja AT ikarustech DOT com
9 years 4 weeks ago #3676

Please Log in or Create an account to join the conversation.

Replied by Jasem Mutlaq on topic Alignment/Plate solve problems

mtoastro, I was able to solve your M51 image in 9 seconds in Ekos using the _latest_ astrometry.net package. I noted something, your astrometry.net options are missing the -O overwrite option:

This is what I used.

I also installed index files 2208-2219 to cover the FOV.
The following user(s) said Thank You: Marko T.
Last edit: 9 years 4 weeks ago by Jasem Mutlaq.
9 years 4 weeks ago #3678

Please Log in or Create an account to join the conversation.

  • Posts: 43
  • Thank you received: 1

Replied by Marko T. on topic Alignment/Plate solve problems

Indeed it works with the parameters you supplied.
I could have sworn that I tried that. It might have been before I recompiled everything.

Thanks a lot.
9 years 4 weeks ago #3680

Please Log in or Create an account to join the conversation.

  • Posts: 47
  • Thank you received: 0

Replied by Steen on topic Alignment/Plate solve problems

It seems my problem has also found a solution. I tried to solve from commandline with:

solve-field -O --scale-low 1 ngc869_Light_04.fits

to be sure to have the right FOV. This solved after several minutes with:

Field 1: solved with index index-4209.fits.
Field 1 solved: writing to file ./ngc869_Light_04.solved to indicate this.
Field: ngc869_Light_04.fits
Field center: (RA,Dec) = (35.01, 57.3) deg.
Field center: (RA H:M:S, Dec D:M:S) = (02:20:02.662, +57:17:56.302).
Field size: 83.63 x 55.7355 arcminutes
Field rotation angle: up is -179.702 degrees E of N
Creating new FITS file "./ngc869_Light_04.new"...

From here I saw that the FOV was 83.63 x 55.74, and in Ekos it had been calculated to 71.2 x 51.3 . If I in Ekos I changed the FOV to that by changing the focal length from 1000 to 950 everythig is working within 5 secs :-). Why I have to change the FL from what it actual is I don't know. I have cheched pix-size and image size and this seems to be correct, and I'm not using a reducer...

OK - by trying to uses the commandline, you can figure out the right FOV if you have the same problems like me.
9 years 4 weeks ago #3706

Please Log in or Create an account to join the conversation.

Time to create page: 2.171 seconds