pgmkernel (1)

NAME

pgmkernel - generate a convolution kernel

SYNOPSIS

pgmkernel [ -weight w ] width [ height ]

DESCRIPTION

Generates a portable graymap array of size width x height (or width x width if height is not specified) to be used as a convolution file by pnmconvol. The data in the convolution array K are computed according to the formula:

K(i,j) = 1 / ( 1 + w * sqrt((i-width/2)^2 + (j-height/2)^2))

where w is a coefficient specified via the -weight flag, and width and height are the X and Y filter sizes.

The output PGM file is always written out in ASCII format.

OPTIONS

The optional -weight flag should be a real number greater than -1. The default value is 6.0.

BUGS

The computation time is proportional to width * height. This increases rapidly with the increase of the kernel size. A better approach could be using a FFT in these cases.

SEE ALSO

pnmconvol(1), pnmsmooth(1)

AUTHOR

Alberto Accomazzi (alberto@cfa.harvard.edu).