fixed Annotation in JNow instead of J2000 - possible ?
- Thread starter Quasarius
- Start date
Hi Thomas,
If that happens with one of our spline-based astrometric solutions, it is because the catalog coordinates and proper motions are wrong or inaccurate at the scale of the image. In other words, you don't need—and should not use—apparent or intermediate coordinates (which are the correct names of what is loosely called 'JNow' for equinox-based and CIO-based positions, respectively) when working with our native astrometric solutions.
Our astrometric solutions are referred to either the ICRS or the GCRS. ICRS solutions use astrometric positions, including corrections for space motion (parallax, radial velocity, and proper motions with corrections for the relativistic Doppler effect) and relativistic light deflection due to solar gravitation. GCRS solutions use proper positions, which include the same transformations as astrometric positions, plus corrections for annual and diurnal aberration of light (relativistic model) to define the actual direction of each source as seen by the observer. Our position reduction and solar system ephemeris calculation routines implement rigorous relativistic models and formulations with sub-milliarcsecond accuracy.
Currently, Gaia is the only stellar catalog able to provide positions and proper motions at the accuracy level of our astrometric solutions. You should always use the Gaia DR3 catalog as local XPSD database files to perform any astrometry tasks in PixInsight, including accurate image annotations.
If that happens with one of our spline-based astrometric solutions, it is because the catalog coordinates and proper motions are wrong or inaccurate at the scale of the image. In other words, you don't need—and should not use—apparent or intermediate coordinates (which are the correct names of what is loosely called 'JNow' for equinox-based and CIO-based positions, respectively) when working with our native astrometric solutions.
Our astrometric solutions are referred to either the ICRS or the GCRS. ICRS solutions use astrometric positions, including corrections for space motion (parallax, radial velocity, and proper motions with corrections for the relativistic Doppler effect) and relativistic light deflection due to solar gravitation. GCRS solutions use proper positions, which include the same transformations as astrometric positions, plus corrections for annual and diurnal aberration of light (relativistic model) to define the actual direction of each source as seen by the observer. Our position reduction and solar system ephemeris calculation routines implement rigorous relativistic models and formulations with sub-milliarcsecond accuracy.
Currently, Gaia is the only stellar catalog able to provide positions and proper motions at the accuracy level of our astrometric solutions. You should always use the Gaia DR3 catalog as local XPSD database files to perform any astrometry tasks in PixInsight, including accurate image annotations.
Quasarius
Active member
Thanks for your answer Juan.
I don't know if I have explained the problem correctly.
Astrometric solution is perfect, also annotation, all objects marked exactly at the right place in image, PGC, QSO's, very nice.
I use of DR3 database.
But, this is only true for distant objects. If I want to mark nearby stars (AnnotateImage GCVS variable stars for instance), markers are not at right place.
It is the same in CdC Cartes du Ciel, if I use J2000 there, the position difference for these stars is exactly the same as in Pixinsight.
And if I tell CdC to use JNow, all is ok, the positions (e.g. for variable stars) are correct again.
And this gave me the idea to ask Pixinsight for a possibility to display/convert from J2000 to JNow for the display/overlay of annotations.
I don't know if I have explained the problem correctly.
Astrometric solution is perfect, also annotation, all objects marked exactly at the right place in image, PGC, QSO's, very nice.
I use of DR3 database.
But, this is only true for distant objects. If I want to mark nearby stars (AnnotateImage GCVS variable stars for instance), markers are not at right place.
It is the same in CdC Cartes du Ciel, if I use J2000 there, the position difference for these stars is exactly the same as in Pixinsight.
And if I tell CdC to use JNow, all is ok, the positions (e.g. for variable stars) are correct again.
And this gave me the idea to ask Pixinsight for a possibility to display/convert from J2000 to JNow for the display/overlay of annotations.
fredvanner
Well-known member
Does the Gaia DR3 XPDS database include proper motion data?
fredvanner
Well-known member
If I understand Juan correctly, all annotation is converted to the epoch of the image. The one thing that is not clear to me is whether the Gaia database includes proper motion data. If it does, and this is used for the corrected positions, then the annotation should be correct...
... but I do know that a relatively small proportion of the Gaia data release has accurate proper motion data, so I don't know if PixInsight chose to include those fields in the XPSD database.
Again, this doesn't make any sense. Annotations of "nearby" stars are as accurate as any other annotations, provided that the correct coordinates are used. The correct coordinates can only be ICRS astrometric positions or GCRS proper positions, depending on the reference system used to compute the astrometric solution.
There are no other valid choices. Using apparent or "of date" places with an ICRS solution is wrong because different coordinate systems are being mixed. ICRS is an inertial reference system, while of date or apparent/intermediate positions are referred to a rotating system tied to Earth's equator and pole, where the effects of frame bias, precession and nutation must be taken into account. You are mixing incompatible coordinates, and if the result is apparently valid, there must be another cause out of our normal control.
cloudbait
Well-known member
Perhaps you could provide one or more examples of objects that are not properly annotated. (And an image where this issue occurs.)
Last edited:
Our XPSD Gaia databases contain the following data for each catalog source:
- Right ascension and declination.
- Parallax.
- Proper motions in right ascension and declination.
- G, BP and RP magnitudes.
- A 32-bit flags word providing information on data quality and availability.
Gaia represents the current state of the art in astrometry and (spectro)photometry by a considerable margin. There is definitely no other choice for accurate astrometry at the milliarcsecond level.
cloudbait
Well-known member
So you're looking for the option to precess the coordinates of a J2000 catalog to the image time, right?
), AD Leo is misplaced (it is the star just under the marker). Image solved with DR3 database, AnnotateImage script and choose QSO's and variable stars to annotate.
Quasarius
Active member
Yes.
Can you please upload the image for testing? We need the linear image where you are observing this problem.
See for example this annotation with a bunch of GCVS stars in the field:
The GCVS catalog cannot be compared with Gaia in terms of positional accuracy, but the annotations are quite good for all stars. The image scale is about 1 arcsecond/pixel.
The GCVS catalog cannot be compared with Gaia in terms of positional accuracy, but the annotations are quite good for all stars. The image scale is about 1 arcsecond/pixel.
This forum has limited storage capabilities. You have to upload the image using a public cloud storage service such as Google Cloud, Dropbox, etc. We need a publicly shared link to the image.
Quasarius
Active member
cloudbait
Well-known member
Hmm. I said "precess", but the issue you're having is actually with proper motion, isn't it? Otherwise distance wouldn't matter. But changing the epoch shouldn't matter in that case.