owin2mask: Convert Window to Binary Mask under Constraints

In spatstat.geom: Geometrical Functionality of the 'spatstat' Family

Convert Window to Binary Mask under Constraints

Description

Converts a window (object of class "owin") to a binary pixel mask, with options to require that the inside, outside, and/or boundary of the window should be completely covered.

Usage

owin2mask(W,
         op = c("sample", "notsample",
                "cover", "inside",
                "uncover", "outside",
                "boundary",
                "majority", "minority"),
         ...)

Arguments

W	A window (object of class "owin").
op	Character string (partially matched) specifying how W should be converted to a binary pixel mask.
...	Optional arguments passed to as.mask to determine the pixel resolution.

Details

This function is similar to, but more flexible than, as.mask. It converts the interior, exterior, or boundary of the window W to a binary pixel mask.

• If op="sample" (the default), the mask consists of all pixels whose centres lie inside the window W. This is the same as using as.mask.

• If op="notsample", the mask consists of all pixels whose centres lie outside the window W. This is the same as using as.mask followed by complement.owin.

• If op="cover", the mask consists of all pixels which overlap the window W, wholly or partially.

• If op="inside", the mask consists of all pixels which lie entirely inside the window W.

• If op="uncover", the mask consists of all pixels which overlap the outside of the window W, wholly or partially.

• If op="outside", the mask consists of all pixels which lie entirely outside the window W.

• If op="boundary", the mask consists of all pixels which overlap the boundary of the window W.

• If op="majority", the mask consists of all pixels in which at least half of the pixel area is covered by the window W.

• If op="minority", the mask consists of all pixels in which less than half of the pixel area is covered by the window W.

These operations are complementary pairs as follows:

"notsample"	is the complement of	"sample"
"uncover"	is the complement of	"inside"
"outside"	is the complement of	"cover"
"minority"	is the complement of	"majority"

They also satisfy the following set relations:

"inside"	is a subset of	"cover"
"outside"	is a subset of	"uncover"
"boundary"	is a subset of	"cover"
"boundary"	is a subset of	"uncover"

The results of "inside", "boundary" and "outside" are disjoint and their union is the entire frame.

Theoretically "sample" should be a subset of "cover", "notsample" should be a subset of "uncover", "inside" should be a subset of "majority" and "outside" should be a subset of "minority", but these could be false due to numerical error in computational geometry.

Value

A mask (object of class "owin" of type "mask" representing a binary pixel mask).

Author(s)

\adrian

.

See Also

as.mask

Examples

  W <- Window(chorley)
  opa <- par(mfrow=c(2,5))
  plot(as.mask(W, dimyx=10), col="grey", main="sample")
  plot(W, add=TRUE, lwd=3, border="red")  
  plot(owin2mask(W, "ma", dimyx=10), col="grey", main="majority")
  plot(W, add=TRUE, lwd=3, border="red")
  plot(owin2mask(W, "i", dimyx=10), col="grey", main="inside")
  plot(W, add=TRUE, lwd=3, border="red")
  plot(owin2mask(W, "c", dimyx=10), col="grey", main="cover")
  plot(W, add=TRUE, lwd=3, border="red")
  plot(owin2mask(W, "b", dimyx=10), col="grey", main="boundary")
  plot(W, add=TRUE, lwd=3, border="red")
  plot(as.mask(complement.owin(W), dimyx=10), col="grey",
       main="notsample")
  plot(W, add=TRUE, lwd=3, border="red")  
  plot(owin2mask(W, "mi", dimyx=10), col="grey", main="minority")
  plot(W, add=TRUE, lwd=3, border="red")
  plot(owin2mask(W, "o", dimyx=10), col="grey", main="outside")
  plot(W, add=TRUE, lwd=3, border="red")
  plot(owin2mask(W, "u", dimyx=10), col="grey", main="uncover")
  plot(W, add=TRUE, lwd=3, border="red")
  plot(owin2mask(W, "b", dimyx=10), col="grey", main="boundary")
  plot(W, add=TRUE, lwd=3, border="red")
  par(opa)

spatstat.geom documentation built on Oct. 20, 2023, 9:06 a.m.