DynamicCrop and CopyCoordinates - Plate Solve Data Issues

[dcarr]

Hello

There are a couple of places where it seems invalid solve data can propagate from image to image.

1   DynamicCrop. I routinely use the ImageSolver and AnnotateImage scripts, and these work very well.  However, if I subsequently use DynamicCrop, the plate solve data appears to be transferred unchanged to the cropped image.  This leads to a cropped image with invalid solve data (The level of wrongness will be worse if rotation is involved).  I suggest two alternative solutions: 1. Simply delete the invalid solve data from the cropped image. Better to have no data than invalid data.  Or 2. Update the solve data from the original image using transforms based on the parameters used for the DynamicCrop.

2   CopyCoordinates. To overcome the issue above, I generally use the CopyCoordinates script (from original to cropped image, with Transform Copy option).  However, a minor issue arises.  The Focal Length and Pixel Size data do not appear to be transferred correctly.  For me, pixel size is 5.4 microns, focal length is 910mm and image scale is about 1.22 arcsec per pixel.  However, it appears the CopyCoordinates script records a default pixel size of 7.4 microns and a calculated value of about 1200 mm for focal length in the target image.  This is of no consequence to me, but may be more so to others.  The data is corrected if/when I run ImageSolver as a confirmatory step after CopyCoordinates. Sometimes, I just manually edit the FITS header to correct the data. 

Let me take the opportunity to compliment you on terrific software.  As you can see, there are simple workarounds to these minor matters. Please let me know if I am making an error here.

Regards
Dean

[Andres.Pozo]

[...]CopyCoordinates. [...]However, a minor issue arises.  The Focal Length and Pixel Size data do not appear to be transferred correctly.  [...]

I'll look into this.

Regarding the propagation of the astrometric metadata when changing the geometry of the images, it is out of my hands. All I can do by myself I already did with CopyCoordinates.

[Juan Conejero]

Regarding the propagation of the astrometric metadata when changing the geometry of the images, it is out of my hands. All I can do by myself I already did with CopyCoordinates.

Today I'm going to release an update to the Geometry module to solve this problem (among other improvements). While we don't have a better solution, all geometric tools changing image geometry will remove existing astrometric solutions. These tools will also add HISTORY keywords to inform about the changes they perform to the image.

Note however that the focal length and pixel size keywords and properties, among others, refer to image acquisition, so they remain valid even after modifying image dimensions an orientation. In other words, with images that have already been processed, never rely on FOCALLEN and XPIXSZ/YPIXSZ keywords to draw conclusions about image scale.

[dcarr]

Thank you Juan and Andres for your very quick responses to these matters.  PixInsight gets better and better.

Yes, Focal Length and Pixel Size are only references to the original image acquisition setup. I sometimes refer to them later as additional reminders about the specific telescope/camera/reducer arrangement used at the time.  There are plenty of places where image scale can change during processing.

Thanks again.

Dean