Help with plate solve

[Duncan]

Don't know if anybody can help with this, but I'm having serious difficulty with a plate solve (using the ImageSolver script).

I've been able to get this to work without problems before, but this image creates some serious problems, I'm assuming because of the high density of stars in the background (image attached*). I can't get the plate solve included in TheSkyX to work on this either.

I was able to get the solver to work ok on one of the source FITS files, and I can get it to work on the final image if I do a histogram transformation to drop out most of the dimmer stars. Just wondering - is there an easy way to copy the solution from one image to another? I could clone the image, do the histo transform and run the solver on the duplicate, but don't know if there's a non-manual way of copying the solution back to the original image's FITS header.

Any ideas?

*this is compressed very heavily to fit the forum rules... the nasty yellow artifacts in the background are JPG artifacts 

[aworonow]

Two things I often try are 1) changing the magnitude limit in the solver dialog and 2) submitting the image to astronometry.net then using its solution as the starting point in PI.

Another method I've recently had success using is to do the image solver in the new photometric calibration process and enter those into the image solver dialog. The new routine seems much more robust.

AlexW

[Duncan]

Many thanks for the reply. I'll try adjusting the star detection parameters. The typical issue is that the algorithm doesn't find sufficient stars; it frequently says that it found 0 stars in many cases.

I already know the center coordinates; I'm trying to get an astrometric solution for annotation purposes. Maybe I should just copy the result over line by line from the FITS header of one of the source files?

[aworonow]

Yeah, I've often spent a goodly amount of time with that script. Also make sure your focal length and pixel size are fairly accurate; I have tried using the "Resolution" input with poor success. There's probably some interaction it has with pixel size. Finally, if all else fails, and if you started from linear images, try solving one of those. There is an option in the annotation script for applying a stretch to the annotated image.

AlexW

[oldwexi]

Hi Duncan, the solution ist very simple with PixelMath.
PixelMath only transfers the Pixels and not the FITs info.
So assuming your final image is labelled "final"
write
 
final

in the upper inputline of PixelMath and drop the triangle over the solved  image.
Assuming the size of both images is equal.
If not, register the target image to the image "final"  before PixelMath

Gerald

[Andres.Pozo]

Another method I've recently had success using is to do the image solver in the new photometric calibration process and enter those into the image solver dialog. The new routine seems much more robust .

PhotometricColorCalibration uses internally the script ImageSolver. PCC provides a different user interface that gives the option of choosing automatically the magnitude of the catalog, but the solving algorithm is shared by both PCC and ImageSolver.

[msmythers]

You can run PCC to plate solve an image which is nice since it is more automated then the standard Image Solver script. If you do not want the color calibration to happen to your image there is a check box for the calibration at the top of the tool. Just uncheck that box. While color calibration will not be applied it to the image color calibration will be calculated during the use of PCC which can take some time depending on the image. If by chance the calibration fails then the image solver part of PCC does not get applied to the image I beleive.


Mike

[Duncan]

Hi Duncan, the solution ist very simple with PixelMath.
PixelMath only transfers the Pixels and not the FITs info.
So assuming your final image is labelled "final"
write
 
final

in the upper inputline of PixelMath and drop the triangle over the solved  image.
Assuming the size of both images is equal.
If not, register the target image to the image "final"  before PixelMath

Gerald

Gerald,
awesome - that sounds like exactly the solution I need!

I'll try that out when I get home, but I think that will work perfectly

Duncan