pcf_directions: Anisotropic pair correlation function, direction vector...

In antiphon/Kdirectional: Analysis of anisotropy in point patterns using second order statistics

Anisotropic pair correlation function, direction vector formulation

Description

Estimate the anisotropic pair correlation function (2d and 3d), as defined in Stoyan 1991, f. 5.2-5.3. This one doesn't use spherical coordinates but direction vectors.

Usage

pcf_directions(
  x,
  directions,
  h,
  f = 0.15,
  correction = "translation",
  n_dir,
  r
)

Arguments

x	pp, list with $x~coordinates $bbox~bounding box
directions	Matrix of direction vectors
h	widths of epanechnicov kernels, vector of two values, for ranges and angles.
f	If h not given, use h=f/lambda^(1/dim) for range and h=f*pi
correction	"none" or translation. Translation only for rectangle box.
n_dir	Direction vector grid resolution (if units not given). See Details.
r	In case directions not given, use these lengths for generated directions.

Details

For example, in 2D the polar vector (r, theta) is now a vector (r*cos(theta), r*sin(theta)).

For 2d, we use a regular grid of n_dir steps on [0,pi], leaving out pi due to antipode-symmetry to 0. n_dir=7 is the default.

For 3d we use a triangular grid on the expanding sphere. This is generated by subdiving a regular icosahedron n_dir times. n_dir=2 is the default.

antiphon/Kdirectional documentation built on Feb. 13, 2023, 6:26 a.m.