RETRO commands
PROGRAM NAME
vmarr - Create a Marr Type Edge Detection Filter Kernel (K1)
DESCRIPTION
creates a convolution filter kernel using the following equation.
pixel = [-1/(PI*var*var)]*[1-r*r/(2*var)]*exp[-r*r/(2*var)],
where r is the calculated distance from the center of the kernel.
A Marr filter is the Laplacian of
a Gaussian, and is intended for high lighting image
edges and at the same time smooth out the noise.
vmarr creates image data of type float.
REQUIRED ARGUMENTS
- -o
-
type: outfile
desc: Marr kernel output image
OPTIONAL ARGUMENTS
- -trigger
-
type: infile
desc: trigger input
default: {none}
- -r
-
type: integer
desc: specify the number of rows in the kernel
default: 3
bounds: value > 0
- -c
-
type: integer
desc: specify the number of columns in the kernel
default: 3
bounds: value > 0
- -v
-
type: float
desc: variance of the Gaussian
default: 1
bounds: value > 0.0
EXAMPLES
SEE ALSO
kasc2val, kimportasc
RESTRICTIONS
only generates float kernels, kconvert can change the
image to other data types.
REFERENCES
COPYRIGHT
Copyright (C) 1993 - 1997, Khoral Research, Inc. ("KRI") All rights reserved.