.TL
Pgmkernel User Manual
.SH 1
pgmkernel
.LP
Updated: 10 December 1992
.br
Table Contents
.SH 2
NAME
.LP
pgmkernel - generate a convolution kernel
.SH 2
SYNOPSIS
.LP
\fBpgmkernel\fR
[\fB-weight\fR \fIw\fR]
\fIwidth \fR
[\fIheight \fR]
.SH 2
DESCRIPTION
.LP
.LP
This program is part of Netpbm.
.LP
\fBpgmkernel\fR generates a convolution kernel that you can use
with \fBpnmconvol\fR.  The kernel is one where the weight of each location
is inversely proportional to its distance from the center of the kernel.
.LP
\fBpgmkernel\fR generates a PGM image of size \fIwidth\fR by
\fIheight\fR (or \fIwidth\fR by \fIwidth\fR if you don't specify
\fIheight\fR).
.LP
\fBpgmkernel\fR computes the convolution function K as follows.
K(i,j) = 1 / ( 1 + w * sqrt(i^2 + j^2)) 
where \fIw\fR is a coefficient specified via the \fB-weight\fR
option.  \fIi\fR and \fIj\fR are measured in pixels.  K is zero
everywhere beyond the specified kernel width and height.
.LP
\fBpgmkernel\fR generates the output PGM file in the Plain (text)
variation of PGM.
.SH 2
OPTIONS
.LP
The optional \fB-weight\fR value should be a real number greater than
-1.  The default value is 6.0.
.SH 2
LIMITATIONS
.LP
.LP
The computation time is proportional to \fIwidth\fR*\fIheight\fR.
This increases rapidly with the increase of the kernel size.  A better
approach could be using a FFT in these cases.
.SH 2
SEE ALSO
.LP
pnmconvol,
pnmsmooth
pamgauss
pgm
.SH 2
AUTHOR
.LP
Alberto Accomazzi (alberto@cfa.harvard.edu).
.br
\l'5i'
.SH 2
Table Of Contents
.LP
.IP \(bu
NAME
.IP \(bu
SYNOPSIS
.IP \(bu
DESCRIPTION
.IP \(bu
OPTIONS
.IP \(bu
BUGS
.IP \(bu
SEE ALSO
.IP \(bu
AUTHOR
.LP