Anisotropy profile for given diagonal transformations

Compares the directed sector-K functions along main axes.

Usage

anisotropy_profile(
  x,
  grid,
  r,
  verb = FALSE,
  eps = pi/4,
  border = TRUE,
  antipodal = TRUE,
  power = 2,
  ...
)

Arguments

x

point pattern

grid

p x dim matrix giving the diagonals of the candidate transformations

r

vector of range values to sum over ('integration nodes')

verb

verbose?

eps

Sector angle

border

Border correction? Should use this unless the data is large.

antipodal

Use antipodally symmetric estimation of sector-K?

power

Power of the differences, absolute value for power=1.

Details

For each v in grid, transform x -> T(v)x and summarize anisotropy by absolute integral difference in x-y-z-directed K-functions.

The transformations T(v) are given by diagonal vector v, rows of parameter grid.

We assume that no rotation is present, or that rotation is already corrected for, so that squared difference sums of the directed sector-K functions along main axes should indicate the optimal v.