spatstat (version 1.63-0)

rotmean: Rotational Average of a Pixel Image

Description

Compute the average pixel value over all rotations of the image about the origin, as a function of distance from the origin.

Usage

rotmean(X, ..., origin, padzero=TRUE, Xname, result=c("fv", "im"))

Arguments

X

A pixel image.

…

Ignored.

origin

Optional. Origin about which the rotations should be performed. Either a numeric vector or a character string as described in the help for shift.owin.

padzero

Logical. If TRUE (the default), the value of X is assumed to be zero outside the window of X. If FALSE, the value of X is taken to be undefined outside the window of X.

Xname

Optional name for X to be used in the function labels.

result

Character string specifying the kind of result required: either a function object or a pixel image.

Value

An object of class "fv" or "im".

Details

This command computes, for each possible distance \(r\), the average pixel value of the pixels lying at distance \(r\) from the origin. Kernel smoothing is used to obtain a smooth function of \(r\).

If result="fv" (the default) the result is a function object of class "fv" giving the mean pixel value of X as a function of distance from the origin.

If result="im" the result is a pixel image, with the same dimensions as X, giving the mean value of X over all pixels lying at the same distance from the origin as the current pixel.

If padzero=TRUE (the default), the value of X is assumed to be zero outside the window of X. The rotational mean at a given distance \(r\) is the average value of the image X over the entire circle of radius \(r\), including zero values outside the window if the circle lies partly outside the window.

If padzero=FALSE, the value of X is taken to be undefined outside the window of X. The rotational mean is the average of the X values over the subset of the circle of radius \(r\) that lies entirely inside the window.

Examples

# NOT RUN {
  if(interactive()) {
    Z <- setcov(square(1))
    plot(rotmean(Z))
    plot(rotmean(Z, result="im"))
  } else {
    Z <- setcov(square(1), dimyx=32)
    f <- rotmean(Z)
  }
# }