fitskypars
salgorithm = mode
The sky fitting algorithm to be employed. DAOPHOT users should
in general use the "mode" algorithm for best results in very
crowded fields. The sky fitting options are:
constant
Use a user supplied constant value for sky.
file
Read sky values from a text file.
mean
Compute the mean of the sky pixel distribution.
median
Compute the median of the sky pixel distribution.
mode
Compute the mode of the sky pixel distribution.
centroid
Compute the intensity weighted mean (centroid) of the sky pixel histogram.
gauss
A Gaussian function is fit to the sky pixel histogram using nonlinear least
squares techniques.
ofilter
A variation of the optimal filtering technique using a triangular weighting
function is applied to the histogram of the sky pixels.
crosscor
The cross-correlation function of the pixel distribution and a Gaussian
noise function is used to compute the sky value.
histplot
The peak of the histogram of sky pixels is marked interactively by the
user on the graphics device.
radplot
The sky level can be marked interactively on a radial profile plot.
annulus = 10.0 (scale units)
The inner radius of the annular sky fitting region in units of the
scale parameter.
dannulus = 10.0 (scale units)
The width of the annular sky fitting region in units of the scale parameter.
skyvalue = 0.0
The constant for constant sky subtraction.
smaxiter = 10
The maximum number of iterations performed by the sky fitting algorithm.
Smaxiter is required by the gauss and ofilter sky fitting algorithms.
sloclip = 0.0, shiclip = 0.0 (percent)
The high and lo side clipping parameters in percent of the total number
of pixels. If either of these parameters > 0.0 then the specified
percentage of the pixels will be arbitrarily removed from the sky pixel
distribution before any sky fitting is done.
snreject = 50
The maximum number of pixel rejection cycles.
sloreject = 3.0, shireject = 3.0
The ksigma clipping factors for the pixel rejection phase of the
sky fitting algorithm. Sloreject and shireject are
in units of the computed sky sigma.
khist = 3.0
The ksigma clipping factor for computing the histogram of the sky pixels.
Khist is in units of
sigma of the local sky pixel distribution.
The histogram will be 2.0 * khist * sigma wide.
binsize = 0.10
The width of a single bin of the histogram of sky values.
Binsize is in units of the sigma of the local sky pixel distribution.
smooth = no
Boxcar smooth the histogram before computing a sky value.
rgrow = 0.0 (scale units)
The region growing radius for pixel rejection in the sky region, in units
of scale. When a bad sky_pixel is detected, all pixels within
rgrow / scale will be rejected. If rgrow is 0.0 region growing
is not performed.
The sky fitting algorithm parameters control the action of the sky fitting algorithms. The default values have been chosen for efficiency and to give reasonable results in the majority of cases. Note that many of the sky fitting parameters scale with two quantities which are data dependent, the image scale, scale, and the standard deviation of the local sky pixels. The scale parameter is located in the DATAPARS task.
Using an initial position and scale supplied by the user, sky pixels in an annular region of inner radius annulus / scale pixels and width dannulus / scale pixels are extracted from the IRAF image. Pixels outside the good data range specified by the DATAPARS task parameters datamin and datamax are excluded from the sky fit. Sloclip and shiclip percent of the pixels are also removed from either end of the sky distribution before any fitting is done.
If the scale parameter is defined in terms of the half width at half maximum of the point spread function, then single values of annulus and dannulus once chosen will work well for for computing sky values in images of different seeing and detector characteristics.
Sky values are computed using the sky fitting algorithm specified by salgorithm. The default value is mode. If fitsky = median or mode, FITSKY computes the the appropriate sky value directly from the array of sky pixels.
The remaining sky fitting algorithms employ the histogram of the object sky pixels. The computed histogram is 2.0 * khist * sigma wide with a bin width of binsize * sigma. where sigma is the computed standard deviation of the sky pixels. If smooth = yes, boxcar smoothing is performed on the computed histogram before sky fitting. The mode of the histogram is then computed by, a non-linear least squares fit to a Gaussian (salgorithm = gauss), optimal filtering of the histogram (salgorithm = ofilter), computing the centroid of the histogram (salgorithm = centroid) or by cross-correlation techniques (salgorithm = crosscor).
Two interactive methods of fitting sky are also available. If salgorithm is radplot or histplot, the user can interactively set the value of the sky using the standard graphics device and graphics cursor.
Pixels which deviate from the sky value by more than sloreject * skysigma and shireject are rejected from the fit. If rgrow > 0, pixels within a radius of rgrow / scale of the rejected pixel are also rejected from the fit. The procedure iterates until no more pixels are rejected, all pixels are rejected or the maximum number of rejection cycles snreject iterations is reached.
1. List the sky fitting parameters.
da> lpar fitskypars
2. Edit the sky fitting parameters.
da> fitskypars
3. Edit the FITSKYPARS parameters from with the PHOT task.
da> epar phot
... edit a few phot parameters
... move to the fitskypars parameter and type :e
... edit the fitskypars parameters and type :q
... finish editing the phot parameters and type ^Z
4. Save the current FITSKYPARS parameter set in a text file skynite1.par. This can also be done from inside a higher level task as in the above example.
da> epar fitskypars
... type ":w skynite1.par" from within epar
TIME REQUIREMENTS