Limiting the field of view

[James Gordon]

Hello,

The plate solver that I compiled locally on Ubuntu seems to ignore the lower and upper bound of the image scale when executed as
solve-field pleiades.jpg -8 neg -u degwidth -L 5 -U 20 -O --downsample 2

I guess I made some mistake because the online solver marks the same file correctly. Just as an additional note, I have the index file that solved the plate online:

35 reference sources within the image.
RoR: 357948
Test stars in RoR: 3885 of 3885
Good bins: 117 / 117; effA 2.4064e+07 of 2.4064e+07
Ref stars in RoR: 35 of 35
Logodds: 305.852
35 matches, 74 distractors, 0 conflicts (at best log-odds); 3885 field sources, 35 index sources
35 matches, 3850 distractors, 0 conflicts (all sources)
Hit/miss: ++++++++-+--++-+++++++--+---++--+---+-----+-----+++---------+-+-+-------+-----------+-----------+--+
Tweak2: final WCS:
Found tag-along columns from field: FLUX BACKGROUND
Field 1: solved with index index-4112.fits.

Locally the verbose output for this index file is:

Index name "/usr/local/astrometry/data/index-4112.fits" is readable; assuming singe file.
Index scale: [120, 170] arcmin, [7200, 10200] arcsec
Index has 150528 quads and 94080 stars
Trying index /usr/local/astrometry/data/index-4112.fits...
Solving field 1.
Solver:
  Arcsec per pix range: 2.99202, 107.713
  Image size: 6016 x 4000
  Quad size range: 400, 0
  Objs: 0, 10
  Parity: 1, flip
  Use_radec? no
  Verify_pix: 1
  Code tol: 0.01
  Dist from quad bonus: yes
  Distractor ratio: 0.25
  Log tune-up threshold: 13.8155
  Log bail threshold: -230.259
  Log stoplooking threshold: inf
  Maxquads 0
  Maxmatches 0
  Set CRPIX? no
  Tweak? yes
    Forward order 2
    Reverse order 2
  Indexes: 1
    /usr/local/astrometry/data/index-4112.fits
  Field: 3885 stars
Quad scale range: [400, 3409.74] pixels
object 1 of 10: 0 quads tried, 0 matched.
object 2 of 10: 0 quads tried, 0 matched.
object 3 of 10: 0 quads tried, 0 matched.
object 4 of 10: 0 quads tried, 0 matched.
object 5 of 10: 1 quads tried, 63 matched.
object 6 of 10: 4 quads tried, 206 matched.
Got a new best match: logodds 4.12849.
Got a new best match: logodds 5.36381.
object 7 of 10: 10 quads tried, 560 matched.
object 8 of 10: 13 quads tried, 750 matched.
object 9 of 10: 20 quads tried, 1167 matched.
object 10 of 10: 38 quads tried, 2107 matched.
Field 1: tried 38 quads, matched 2107 codes.
Field 1 did not solve (index index-4112.fits, field objects 1-10).
Best odds encountered: 213.538
Spent 0.392 s user, 0.004 s system, 0.396 s total, 0.408819 s wall time.


Then the solver proceeds to the rest of the index files, and goes beyond the upper limit but never completes.

 Field 1 did not solve (index index-4219.fits, field objects 31-40).

It reaches field objects 51-60 unless I kill the job. I appreciate your help. Thank you very much.

Cheers,
James

[Dustin Lang]

The upper-bound command-line flag is "-H" ("high"), not "-U".

  -L / --scale-low <scale>: lower bound of image scale estimate
  -H / --scale-high <scale>: upper bound of image scale estimate

cheers,
--dustin

[James Gordon]

Oh, thank you. What a silly mistake of mine.

Maybe there is still something wrong because if executed it in this way:
/usr/local/astrometry/bin/solve-field Seven\ sisters.jpg -8 neg -u degwidth -L 5 -H 20 -O --downsample 2

the field is not solved with the index file that the online solver does, and goes beyond 20 degrees, such as

Field 1 did not solve (index index-4217.fits, field objects 31-40).

Cheers,
James

[Dustin Lang]

I'm not sure what you mean by "goes beyond 20 degrees"... index 4217 contains features up to 16 degrees, which could be contained in an image of 20 degrees in width.

If you want to post a link to your online submission, or the image, we can take a look.

cheers,
--dustin

[James Gordon]

Here is the link. Thank you for the help.

http://nova.astrometry.net/status/1398364

Cheers,
James