Deep Sky Image Annotation slightly off

[vichman]

Hello all,

when using the absolutely cool feature Deep Sky Image Annotation I have run into a weird phenomenon:
When annotating stars, the the annotation is sometimes spot on, sometimes significantly off.

See images below (zoomed in):

Adulfin's annotation marker is spot on

SpotOn.jpg (886.63 KiB) Viewed 1634 times

While for Altair it is significantly off:

Off.jpg (999.23 KiB) Viewed 1634 times

For a bright star this is quite obvious, but for faint objects one cannot be sure whether the marker is correct.

I did nothing different between both images, both were taken with the same setup only some minutes apart (2000mm focal length, full frame camera, plate solving with ASTAP, latest version, D50 database) with the latest version of SharpCap (4.1.11003).
For the annotation I simply used the built-in database of SharpCap, nothing fancy.

Does anyone have an idea what the reason could be?

CS
Stefan

[admin]

Hi,

there are a few reasons this can be slightly off...

One is a mismatch between J2000 and JNOW, but that would be constant for all stars in the same part of the sky.

Another is that SharpCap only uses a linear model for how the RA/Dec co-ordinates change across the image - that can be inaccurate with a large field of view or if you are looking at something near the pole. It's worse if there has been an offset of the mount between the position where the plate solve happened and the position where you are looking at now - again the maths for that offset is only handled to a 1st order approximation.

To keep errors to a minimum, do a new plate solve if you move to a nearby location (or after a 'plate solve and sync' - which moves the mount after the original plate solve to correct the pointing error)

cheers,

Robin

[vichman]

Hi Robin,

thanks for your message, but I would rule out all the reasons you mentioned below.

Hi,

there are a few reasons this can be slightly off...

One is a mismatch between J2000 and JNOW, but that would be constant for all stars in the same part of the sky.

Indeed, this would also be my understanding that in that case the error would be constant for all objects. This is not the case.
In addition I tried all variations of settings for J2000 and JNOW in the mount, drivers, etc., no effect.

Another is that SharpCap only uses a linear model for how the RA/Dec co-ordinates change across the image - that can be inaccurate with a large field of view or if you are looking at something near the pole. It's worse if there has been an offset of the mount between the position where the plate solve happened and the position where you are looking at now - again the maths for that offset is only handled to a 1st order approximation.

To keep errors to a minimum, do a new plate solve if you move to a nearby location (or after a 'plate solve and sync' - which moves the mount after the original plate solve to correct the pointing error)

The mount I use is a high precision 10micron mount, so the pointing error should be very small.
But I understand that the pointing error would not have any effect anyway, if you plate solve the image and do not move the mount afterwards.
I understand that the annotation is related to the image, not to the position of the mount.

I did exactly what you proposed, I did a near plate solve and did not move the mount any more afterwards.
The object is far from the pole, and also the object is in the center of the image, i.e. the size of the field of view (that is rather small anyway) should not matter.

Any other ideas?

One idea: Do you know the coordinates of the star Altair that SharpCap uses in its internal database for the annotation?
Maybe there is a mismatch in the coordinates?
Is there a way I can manually access the list of objects?

CS,
Stefan

[admin]

Hi Stefan,

the co-ordinates of Altair were pretty much the first thing that I checked - the ones SharpCap uses are

RA=19.846388 Dec=8.868322

These checked out to within an arc second or less with the ones published on Wikipedia, so I assumed they were good.

Do you have a captured image of that field of view that I could test with myself to see if I can work anything out.

cheers,

Robin

[vichman]

Hi Robin,

the co-ordinates of Altair were pretty much the first thing that I checked - the ones SharpCap uses are

RA=19.846388 Dec=8.868322

These checked out to within an arc second or less with the ones published on Wikipedia, so I assumed they were good.

I double checked with SIMBAD, and there they list the same coordinates. So they should be fine.
As they are all J2000, how do you transform them into JNOW?

Do you have a captured image of that field of view that I could test with myself to see if I can work anything out.

I added an image of the complete field (had to convert it to JPG to get it under the 1MB limit, so I hope you still get all the info you need).

2023-09-03-1920_5-HIP 97649 (alpha Aql, 53 Aql, Altair)_00001_WithAnnotations.jpg (831.24 KiB) Viewed 1575 times

Thanks
Stefan

[admin]

Hi Stefan,

well, it slowly gets stranger...

The JPG image does solve for me and I see the same offset in my results. I also tried using AstroTortilla as an alternative plate solver and see a similar effect.

I then uploaded the image to the online astrometry.net solving engine and that also seems to place the annotation off center from the star (harder to see - can't seem to get a higher resolution plot) : https://nova.astrometry.net/user_images ... #annotated

J2000 to/from JNOW conversion is handled by the ASCOM library, which should be solid, but anyway should not be required in this case as the star database and the plate solve results are all J2000.

In this case, could it be the proper motion of the star since the J2000 epoch - apparently has 0.54 arcsec/year in RA and 0.39 in Dec, which would add up over 23 years?

cheers,

Robin

[vichman]

Hi Robin,

I ran platesolve and annotate in PixInsight.
For plate solving I used Gaia3, and for annotation "Named Stars" (white), "Tycho2" (green), Hipparcos" (red) and "UCAC3" (blue), all data comes from VizieR. Unfortunately Altair seems to be not in Tycho2 and UCAC3, only in Named Stars and Hipparcos.

Render.jpg (780.42 KiB) Viewed 1540 times

RenderZoom.jpg (277.73 KiB) Viewed 1540 times

The result is even more weird, both are neither spot on nor where SharpCap locates the star (yellow).

I do not think that proper motion would be the reason, as the result would be an offset of 23*0.54=12" in RA and 23*0.39=9" in DE, while the offset between SharpCap's location and the "real" location is about RA 1" and about DE 6".

Very weird.

Stefan

[admin]

Hi Stefan,

looks like nothing is quite managing to get 'pixel perfect' positions - the one I got with Astrotortilla was different to the one I got with Astap...

I don't think the RA is too far off for proper motion - you have to remember that 1" difference in RA is actually 15 seconds of arc due to RA being measured in hours, not degrees.

cheers,

Robin