blob: cb4dc3de396489971120261e082bf90bc21817a0 (
plain) (
blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
|
.help overview Apr98 noao.imred.crutil
.ce
\fBThe Cosmic Ray Package: CRUTIL\fR
The cosmic ray package provides tools for identifying and removing cosmic
rays in images. The tasks are:
.nf
cosmicrays - Remove cosmic rays using flux ratio algorithm
craverage - Detect CRs against average and avoid objects
crcombine - Combine multiple exposures to eliminate cosmic rays
credit - Interactively edit cosmic rays using an image display
crfix - Fix cosmic rays in images using cosmic ray masks
crgrow - Grow cosmic rays in cosmic ray masks
crmedian - Detect and replace cosmic rays with median filter
crnebula - Detect and replace cosmic rays in nebular data
.fi
The best way to remove cosmic rays is using multiple exposures of the same
field. When this is done the task \fBcrcombine\fR is used to combine the
exposures into a final single image with cosmic rays removed. The images
are scaled (if necessary) to a common data level either by multiplicative
scaling, an additive background offset, or some combination of both.
Cosmic rays are then found as pixels which differ by some statistical
amount away for the average or median of the data.
A median is the simplest way to remove cosmic rays. This is an option
with \fBcrcombine\fR. But this does not make optimal use of the data.
An average of the pixels remaining after some rejection operation is better.
If the noise characteristics of the data can be described by a gain and
read noise then cosmic rays can be optimally rejected using the
"crreject" algorithm. This works on two or more images. There are
a number of other rejection algorithms which can be used as described in
the task help.
The rest of the tasks in the package are used when only a single exposure
is available. These include interactive editing with \fBcredit\fR. The
replacement algorithms in this task may also be used non-interactively if
you have a list of pixel coordinates as input. Other tasks automatically
identifying pixels which are significantly higher than surrounding pixels.
The simplest of these tasks is \fBcrmedian\fR. This replaces
cosmic rays with a median value and produces a cosmic ray
mask which is a simple type of integer image where good pixels have a value
of zero and bad pixels have a non-zero value. The tasks \fBcrgrow\fR and
\fBcrfix\fR are provided to use this type of cosmic ray mask. The former
will flag additional pixels within some radius of the flagged pixels in the
mask. The latter is the basic tool for replacing the identified pixels in
the data by neighboring data. It uses linear interpolation along lines or
columns. The median task is simple but it often will flag the cores of
stars or other small but real features.
The task \fBcraverage\fR is similar to \fBcrmedian\fR in that it compares
the pixel values against a smoothed version. Instead of a median it uses
an average with the central pixel excluded. It is more sophisticated
in that it also compares the average against a larger median to see if
the region corresponds to an object. Thus it can detect objects and
the task could be used as a simple object detection task in its own right.
Because the hardest part of cosmic ray detection from a single image is
avoiding truncation of the cores of stars this task does not allow cosmic
rays to be detected where it thinks there is an object. This task is
also more versatile in allow separate mask values and works on a list
of images.
Somewhat more sophisticated algorithms are available in the tasks
\fBcosmicrays\fR and \fBcrnebula\fR. These attempt to determine if a
deviant pixel is the core of a star or part of a linear nebular feature
respectively.
The best use of these tasks is to experiment and iterate. In particular,
one may want to iterate a task several times and use both \fBcosmicrays\fR
and \fBcraverage\fR.
Good hunting!
.endhelp
|
