psf image
Lientjie wrote on Sep 15, 2008
Hi there
I used the daophot task "psf" to compute a psf for my crowded starfield. I selected by hand 10 stars that were isolated and which had average fwhm values. I then generated the image.mag file with phot and the image.pst file by hand using the information in the image.mag file.
The output image "image.psf.fits" I got, has a funny dark shape in the middle although my stars had close to perfect Gaussian shapes. This is bothering me. Isn't this image that "psf" generates supposed to be like a top-view of the light curve and therefore must be the brightest in the centre with a circular shape? I also did the tutorial example of the spiral galaxy with the name dev&ypix.fits as described in the example of the help function of "psf" and got the same funny s-like shape inside my psf image.
Is that correct and can I relax about it? I just want to make sure that when I subtract this psf image, it won't leave a funny trace on my original image.
Thanks
Lientjie
I used the daophot task "psf" to compute a psf for my crowded starfield. I selected by hand 10 stars that were isolated and which had average fwhm values. I then generated the image.mag file with phot and the image.pst file by hand using the information in the image.mag file.
The output image "image.psf.fits" I got, has a funny dark shape in the middle although my stars had close to perfect Gaussian shapes. This is bothering me. Isn't this image that "psf" generates supposed to be like a top-view of the light curve and therefore must be the brightest in the centre with a circular shape? I also did the tutorial example of the spiral galaxy with the name dev&ypix.fits as described in the example of the help function of "psf" and got the same funny s-like shape inside my psf image.
Is that correct and can I relax about it? I just want to make sure that when I subtract this psf image, it won't leave a funny trace on my original image.
Thanks
Lientjie
Francisco Valdes wrote on Sep 15, 2008
Hello Lientjie,
You may have figured this out since this reply is rather late. The DAO PSF model consists of an analytic function (e.g. a gaussian) and an empircal residual image. The .psf file header has information about the analytic function and the image raster has the residuals. It is very normal to see the pattern you describe as well as a dark ring with a bright peak. To see the actual PSF you should run the task SEEPSF which creates a PSF image you can display.
Yours,
Frank Valdes
You may have figured this out since this reply is rather late. The DAO PSF model consists of an analytic function (e.g. a gaussian) and an empircal residual image. The .psf file header has information about the analytic function and the image raster has the residuals. It is very normal to see the pattern you describe as well as a dark ring with a bright peak. To see the actual PSF you should run the task SEEPSF which creates a PSF image you can display.
Yours,
Frank Valdes
Johan wrote on Sep 15, 2008
valdes
Hello Lientjie,
You may have figured this out since this reply is rather late. The DAO PSF model consists of an analytic function (e.g. a gaussian) and an empircal residual image. The .psf file header has information about the analytic function and the image raster has the residuals. It is very normal to see the pattern you describe as well as a dark ring with a bright peak. To see the actual PSF you should run the task SEEPSF which creates a PSF image you can display.
Yours,
Frank Valdes
Frank,
Yes, we figures that out! Thanks.
To what extent should the image where the psf is subtracted really show residuals at the positions of the stars. After subtraction psf we get at the positions of some stars ( a non-neglible number, I would say) where it is clear that the psf is not a perfect fit. Is that normal? Obviously it is not possible to get a real perfect fit.
Johan
Francisco Valdes wrote on Sep 15, 2008
Johan,
Getting a good PSF and subtraction is sometimes as much an art as science. It depends on the data, whether there is a position dependent PSF variation (e.g. distortions) and maybe the right selection of PSF template stars. So it is normal to get some problems such as you describe. The only thing you can do is iterate until you get the best result you can.
Yours,
Frank
Getting a good PSF and subtraction is sometimes as much an art as science. It depends on the data, whether there is a position dependent PSF variation (e.g. distortions) and maybe the right selection of PSF template stars. So it is normal to get some problems such as you describe. The only thing you can do is iterate until you get the best result you can.
Yours,
Frank
Johan wrote on Sep 15, 2008
Frank,
Okay, understand.
Thanks a lot.
Johan
Okay, understand.
Thanks a lot.
Johan
Last post on Sep 15, 2008