eheinrich / pythonphot

Some of the IDL AstroLib photometry algorithms translated to Python. Please email me with bugs, suggestions, omissions, etc.

PythonPhot PSF Fitting Photometry Tutorial

getpsf.py : Generates a point-spread function (PSF) from observed stars at specified locations. Uses the family of "peak fit" modules (pkfit, pkfit_noise, pkfit_norecent, etc) to fit a gaussian to each star and define an array of non-gaussian psf residuals. Returns a 5-element vector defining the gaussian, a 2-d array of psf residuals, and the magnitude of the psf. Also writes out the psf model to a fits file with the gaussian parameters in the header and the residuals in the data array.

rdpsf.py : Read the .fits file created by getpsf.py that contains the psf model gaussian parameters and 2-d array of residuals.

pkfit.py : fit a psf model to an isolated point source pkfit_noise : fitting with an input noise image pkfit_norecent : forced photometry (fitting a peak without recentering) pkfit_norecent_noise : forced photometry with an input noise image

---

EXAMPLE A : Make a psf model

     import getpsf
     import aper
     import numpy as np
     # load FITS image and specify PSF star coordinates
     image = pyfits.getdata(fits_filename)
     xpos,ypos = np.array([1450,1400]),np.array([1550,1600])

     # run aper to get mags and sky values for specified coords
     mag,magerr,flux,fluxerr,sky,skyerr,badflag,outstr = \
            aper.aper(image,xpos,ypos,phpadu=1,apr=5,zeropoint=25,
            skyrad=[40,50],badpix=[-12000,60000],exact=True)

     # use the stars at those coords to generate a PSF model
     gauss,psf,psfmag = \
            getpsf.getpsf(image,xpos,ypos,
                          mag,sky,1,1,np.arange(len(xpos)),
                          5,'output_psf.fits')

---

EXAMPLE B : fit a psf to isolated stars

 import pyfits
 from PythonPhot import pkfit

 # read in the fits images containing the target sources
 image = pyfits.getdata(fits_filename)
 noiseim = pyfits.getdata(fits_noise_filename)
 maskim = pyfits.getdata(fits_mask_filename)

 # read in the fits image containing the PSF (gaussian model
 # parameters and 2-d residuals array.
 psf = pyfits.getdata(psf_filename)
 hpsf = pyfits.getheader(psf_filename)
 gauss = [hpsf['GAUSS1'],hpsf['GAUSS2'],hpsf['GAUSS3'],hpsf['GAUSS4'],hpsf['GAUSS5']]

 # x and y points for PSF fitting
 xpos,ypos = np.array([1450,1400]),np.array([1550,1600])

 # run 'aper' on x,y coords to get sky values
 mag,magerr,flux,fluxerr,sky,skyerr,badflag,outstr = \
          aper.aper(image,xpos,ypos,phpadu=1,apr=5,zeropoint=25,
          skyrad=[40,50],badpix=[-12000,60000],exact=True)

 # load the pkfit class
 pk = pkfit.pkfit_class(image,gauss,psf,1,1,noiseim,maskim)

 # do the PSF fitting
 for x,y,s in zip(xpos,ypos,sky):
      errmag,chi,sharp,niter,scale = \
          pk.pkfit_norecent_noise(1,x,y,s,5)
      flux = scale*10**(0.4*(25.-hpsf['PSFMAG']))
      dflux = errmag*10**(0.4*(25.-hpsf['PSFMAG']))
      print('PSF fit to coords %.2f,%.2f gives flux %s +/- %s'%(x,y,flux,dflux))