Description:
Sets
to bad, all the pixels with a count above the threshold level. An option allows the close neighbours of
the bright pixel to be set to bad as well. Close neighbours are considered to be those pixels within a
user defined radius of the bright pixel.
A further option allows all bad pixels in the output image to be assigned a random value. The values
chosen are taken from a Normal distribution defined by the user.
Usage:
TOPPED IN OUT WIDTH
BACK SIGMA NSIGMA NOISE PSIZE
Parameters:
BACK = _REAL (Read)
The background count
value. Units counts.
IN = _NDF (Read)
The name of the NDF that is to be examined.
NOISE =
_LOGICAL (Read)
Defines whether or not bad pixels should eventually be assigned a random value.
NSIGMA = _REAL (Read)
The number of standard deviations above sky at which the
cutoff occurs.
OUT = _NDF (Write)
The name of the output NDF that will be created.
PSIZE = _REAL (Read)
The size of each pixel in arc seconds. If the image contains a SKY
co-ordinate frame this value will be determined automatically.
SIGMA = _REAL (Read)
The background pixel count standard deviation value. Units counts.
WIDTH = _REAL
(Read)
The width of the circle around a bright pixel within which pixels will be set to bad.
Units arc seconds.
Examples:
topped in=eggs out=scrambled width=2.5 psize=0.44
back=1000.
sigma=23. nsigma=8. noise=false
Uses EGGS as the input image
and finds all pixels within the image that have a count greater than
1000.8.x23.
These are all set to the bad value. In addition, all pixels within a radius of 1.25 arc seconds are also set
to bad. topped in=objects out=cut width=4. back=6200. sigma=390.
nsigma=10. noise=true
Uses OBJECTS as the input image and
finds all pixels within the image that have a count value greater than
6200.10.x390..
These are all set to random values, as are all the pixels within a radius of 2. arc seconds. Pixel size in
arc seconds will be determined if possible from the WCS component of the image. Notes:
The
distribution of pixel values used when NOISE=TRUE comes from a Normal (Gaussian) distribution.
In some circumstances, particularly for low count values, this may not be appropriate.