daophot.tex

basic median filter simulation

\documentstyle[12pt,titlepage]{article}
\topmargin -0.2in
\oddsidemargin-0.15in
\evensidemargin-0.25in
\textheight 8.5in     
\textwidth 6.5in
\newcommand{\exbegin}{\par\medskip}
\newcommand{\exend}{\medskip\noindent}
\newcommand{\exc}[2]{
\hbox to \hsize{\small\hskip .2in
\parbox[t]{2.2in}{\raggedright\setlength{\parindent}{-.2in}\tt #1}
\hspace{.2in}
\parbox[t]{3.4in}{\raggedright\setlength{\parindent}{-.2in}\rm #2}\hss}
\prevdepth=1.5pt\relax}
% One line example
\newcommand{\exone}[1]{\begin{center}\tt #1 \end{center}}
\begin{document}
\title{IDL and DAOPHOT \\ CCD Version }
\author{W.B. Landsman \\ ST Systems Co.}
\date{9 May 1996}
\maketitle
\section{INTRODUCTION}
This document describes an implementation of an early (1984) version of
the DAOPHOT point-source
photometry algorithm into IDL.  Extensive discussions of DAOPHOT are given   
by Peter Stetson both in his April, 1987 P.A.S.P article, as well as in 
the DAOPHOT user's
manual.
This document will
concentrate on the features of the IDL code that differ from standard
DAOPHOT.   

It should be emphasized that the creators of DAOPHOT have no responsibility
whatsoever for the IDL code.  If you want to use ``true'' DAOPHOT, then use
the FORTRAN version.

We have converted the DAOPHOT algorithms into IDL for two main reasons.
\begin{itemize}
\item The IDL version of DAOPHOT can easily be integrated with the IDL 
TV display and plotting package.  The IDL-DAOPHOT procedures can be 
interrupted at any time to print or plot any variable.
In addition, the use of FITS tables allows 
easy transport of the results to other systems. 

\item The IDL-DAOPHOT code can be easily modified or cannibalized 
for other uses.  For example, the Ultraviolet Imaging Telescope (UIT) group 
has modified the IDL code for use with digitized photographic images. 
The procedures MMM and APER have been adapted for use with extended-source 
photometry.  The procedure GETPSF has been used to supply a point
spread function for deconvolution algorithms.    
\end{itemize}

It should also be emphasized that the IDL code was adapted from an early
version of DAOPHOT, and does not include the following recent improvements
to the DAOPHOT code:

\begin{itemize}
\item FIND will ignore bad pixels when convolving the image with
           a Gaussian PSF. 

\item GETPSF will weight stars by their magnitude before combining
           them to create a final PSF.
      
\item A magnitude dependent fitting radius is used in GROUP, allowing
           stars to be broken up into smaller groups.

\item The most recent versions of DAOPHOT allow a spatially dependent
           PSF.

\item The functions of GROUP and NSTAR have been combined into a single
routine named ALLSTAR.  This presumably speeds up the NSTAR calculation
by reevaluating group membership during NSTAR iterations.

\item The early version of DAOPHOT (and the IDL code) could only represent
the PSF as a Gaussian plus residuals.    More recent versions of DAOPHOT
allow the use of other functions besides a Gaussian.

\end{itemize}

There are 4 ways to obtain the output to the IDL-DAOPHOT procedures.  All
the procedures will display their results at the terminal.    The keyword
/PRINT can be set to also write these results to a file with a default name
(e.g.\ aper.prt), or one can set PRINT = `filename' to specify the name of
the output file. 
By supplying sufficient parameters to a procedure, one can
also store the results in IDL output variables.  Finally, a set of IDL
driver procedures exist, beginning with ``T\_'', which perform I/O to FITS 
tables.  For example, the procedure T\_FIND performs the same results as
FIND, but writes the results to a disk FITS ASCII table.     This file can
be read with any FITS table reader, such as the MRDFITS() function in IDL.
  
A typical DAOPHOT sequence would be to run procedures in the following
order.
\begin{description}
\item[SKY] Determine overall sky background for the image.  This value is
needed to help determine a threshold for FIND.
\item[FIND] Find the positive brightness perturbations (i.e.\ stars) that
meet specified sharpness, roundness, and centroid criteria.
\item[APER] Obtain integrated fluxes within circular apertures for source
positions obtained by FIND.  A sky value for each source is also determined,
using pixels within an annulus of specified inner and outer radii.  This
is the only place in the DAOPHOT sequence where sky values are
determined for individual stars.
\end{description}
APER will give correct results, of course, only when the PSF of individual
stars do not overlap.
\begin{description}
\item[GETPSF]  Obtain a point-spread function (PSF) from one or the sum of
several isolated stars.   The PSF is parameterized as a 2-dimensional
gaussian (integrated over each pixel) and a lookup table of residuals.
If desired, the PSF disk file created by GETPSF can be read using the
procedure RDPSF.
\item[GROUP]  Place stars with non overlapping point-spread functions
into distinct groups.
\item[NSTAR]  Simultaneous point-spread function fitting of all stars 
within each group.   In a sense, the purpose of all the previous 
procedures is to obtain the initial conditions to the NSTAR least-squares
fit.   NSTAR will eliminate stars found by FIND that are too faint, or
that merge with other stars.  However, NSTAR cannot add any stars that
were not found by FIND
\item[SUBSTAR]  Take a set of star positions and magnitudes, (usually from
NSTAR), scale and shift the PSF according to each position and magnitude,
and subtract it from the original image.  The subtracted image can be
used to determine the accuracy of the NSTAR fits, or to search for faint
stars that were missed by FIND (e.g.\ because they were hidden in the wings
of a much brighter star.)  In principle, one could run FIND on the subtracted
image, to obtain additional source positions.
\end{description}
\section{Specific Procedures}
This section briefly describes the use of the IDL-DAOPHOT procedures.
The calling sequence is displayed for both the FITS table, as well as
the direct procedure call.  
Optional parameters are placed within square brackets.  
More detailed information on each procedure can be found in the 
documentation header for the individual procedures.   
For many procedures one can set /DEBUG to print out more diagnostics during
execution.

The first step is to read an image and header (e.g.\ with READFITS)
into IDL variables, which we shall call here IMAGE and IMG\_HDR.  
(An image header is needed only for the FITS table calling sequence.)
\subsection{SKY}
{\tt SKY,IMAGE,[SKYMODE,SKYSIG]} \\

Obtains the overall skylevel of an image by applying the procedure MMM
to approximately 4000 uniformly spaced pixels.  MMM (Mean, Median, Mode)
is sophisticated sky background algorithm that assumes that the sky 
contamination (i.e.\ stars) gives predominantly {\em positive} deviations
from the true value.
The only need for the sky intensity in subsequent 
procedures is to 
determine the threshold detection limit in FIND.
\subsection{FIND}
{\tt FIND,IMAGE,[X,Y,FLUX,SHARP,ROUND,HMIN,FWHM, ROUNDLIM, SHARPLIM]} \\
{\tt T\_FIND,IMAGE,IMG\_HDR, FITSFILE}  \\ 
   
FIND will output the X and Y centroids, an appoximate flux (in magnitudes
relative to the threshold intensity)
and the sharpness and roundness index, for all local enhancements meeting
the threshold, sharpness, roundness, and centroid criteria.  Default
sharpness (0.2-1.0) and roundness (-1.0 - 1.0) limits are used if not
supplied.  FIND will print the number of sources rejected by each criterion,