https://github.com/cran/spatstat
Tip revision: ad0cdcc7b4346a4492f3f16c48c26a8a6b710c8a authored by Adrian Baddeley on 09 May 2014, 11:24:31 UTC
version 1.37-0
version 1.37-0
Tip revision: ad0cdcc
convolve.im.Rd
\name{convolve.im}
\alias{convolve.im}
\title{Convolution of Pixel Images}
\description{
Computes the convolution of two pixel images.
}
\usage{
convolve.im(X, Y=X, \dots, reflectX=FALSE, reflectY=FALSE)
}
\arguments{
\item{X}{
A pixel image (object of class \code{"im"}.
}
\item{Y}{
Optional. Another pixel image.
}
\item{\dots}{Ignored.}
\item{reflectX,reflectY}{
Logical values specifying whether the images \code{X} and \code{Y}
(respectively) should be reflected in the origin before computing
the convolution.
}
}
\value{
A pixel image (an object of class \code{"im"}) representing the
convolution of \code{X} and \code{Y}.
}
\details{
The \emph{convolution} of two pixel images \eqn{X} and \eqn{Y} in the plane
is the function \eqn{C(v)} defined for each vector \eqn{v} as
\deqn{
C(v) = \int X(u)Y(v-u)\, {\rm d}u
}{
C(v) = integral of X(u) * Y(v-u) du
}
where the integral is
over all spatial locations \eqn{u}, and where \eqn{X(u)} and
\eqn{Y(u)} denote the pixel values of \eqn{X} and \eqn{Y} respectively
at location \eqn{u}.
This command computes a discretised approximation to
the convolution, using the Fast Fourier Transform.
The return value is
another pixel image (object of class \code{"im"}) whose greyscale values
are values of the convolution.
If \code{reflectX = TRUE} then the pixel image \code{X} is reflected
in the origin (see \code{\link{reflect}})
before the convolution is computed, so that
\code{convolve.im(X,Y,reflectX=TRUE)} is mathematically
equivalent to \code{convolve.im(reflect(X), Y)}. (These two commands
are not exactly equivalent, because the reflection is performed
in the Fourier domain in the first command, and reflection is
performed in the spatial domain in the second command).
Similarly if \code{reflectY = TRUE} then the pixel image \code{Y} is reflected
in the origin before the convolution is computed, so that
\code{convolve.im(X,Y,reflectY=TRUE)} is mathematically
equivalent to \code{convolve.im(X, reflect(Y))}.
}
\seealso{
\code{\link{imcov}},
\code{\link{reflect}}
}
\examples{
X <- as.im(letterR)
Y <- as.im(square(1))
plot(convolve.im(X, Y))
plot(convolve.im(X, Y, reflectX=TRUE))
plot(convolve.im(X))
}
\author{Adrian Baddeley
\email{Adrian.Baddeley@uwa.edu.au}
\url{http://www.maths.uwa.edu.au/~adrian/}
and Rolf Turner
\email{r.turner@auckland.ac.nz}
}
\keyword{spatial}
\keyword{math}