This application forms a smoothed signal for a one-dimensional NDF whose elements have
oscillating biases. It averages the signal levels of alternating pixels. Both elements
must be good and not deviate from each other by more than a threshold for the averaging
to take place.
This application is intended for images exhibiting alternating patterns in columns or
rows, the so called odd-even noise, arising from electronic interference or readout
through different channels. However, you must supply a vector collapsed along the
unaffected axis, such that the vector exhibits the pattern. See task COLLAPSE
using the Mode or Median estimators. The smoothed image is then subtracted from
the supplied vector to yield the odd-even pattern.
oddeven in out
IN = NDF (Read)
The one-dimensional NDF containing
the input array to be filtered.
OUT = NDF (Write)
The NDF to contain the
THRESH = _DOUBLE (Read)
The maximum difference between
adjacent elements for the averaging filter to be applied. This allows anomalous
pixels to be excluded. If null,
!, is given, then there is no limit.
TITLE = LITERAL (Read)
The title of the output NDF. A null (
!) value means
using the title of the input NDF.
oddeven raw clean The
one-dimensional NDF called raw is filtered such that adjacent pixels are averaged to
form the output NDF call clean.
oddeven out=clean in=raw thresh=20 As
above except only those adjacent pixels whose values differ by no more than
20 are averaged.