Generate a matrix describing the amount of shared boundary length between different planning units, and the total amount of boundary length for each planning unit.

## Usage

```
boundary_matrix(x, ...)
# S3 method for Raster
boundary_matrix(x, ...)
# S3 method for SpatRaster
boundary_matrix(x, ...)
# S3 method for SpatialPolygons
boundary_matrix(x, ...)
# S3 method for SpatialLines
boundary_matrix(x, ...)
# S3 method for SpatialPoints
boundary_matrix(x, ...)
# S3 method for sf
boundary_matrix(x, ...)
# S3 method for default
boundary_matrix(x, ...)
```

## Arguments

- x
`terra::rast()`

or`sf::sf()`

object representing planning units.- ...
not used.

## Value

A `Matrix::dsCMatrix`

symmetric sparse matrix object.
Each row and column represents a planning unit.
Cell values indicate the shared boundary length between different pairs
of planning units. Values along the matrix diagonal indicate the
total perimeter associated with each planning unit.

## Details

This function assumes the data are in a coordinate
system where Euclidean distances accurately describe the proximity
between two points on the earth. Thus spatial data in a
longitude/latitude coordinate system (i.e.,
WGS84)
should be reprojected to another coordinate system before using this
function. Note that for `terra::rast()`

objects
boundaries are missing for cells that have missing (`NA`

) values in all
cells.

## Notes

In earlier versions, this function had an extra `str_tree`

parameter
that could be used to leverage STR query trees to speed up processing
for planning units in vector format.
Although this functionality improved performance, it was not
enabled by default because the underlying function
(i.e., `rgeos:gUnarySTRtreeQuery()`

) was documented as experimental.
The `boundary_matrix()`

function has since been updated so that it will
use STR query trees to speed up processing for planning units in vector
format (using `terra::sharedPaths()`

).

Also, note that in previous versions, cell values along the matrix diagonal indicated the perimeter associated with planning units that did not contain any neighbors. This has now changed such that values along the diagonal now correspond to the total perimeter associated with each planning unit.

## See also

Boundary matrix data might need rescaling to improve optimization
performance, see `rescale_matrix()`

to perform these calculations.

## Examples

```
# \dontrun{
# load data
sim_pu_raster <- get_sim_pu_raster()
sim_pu_polygons <- get_sim_pu_polygons()
# subset data to reduce processing time
r <- terra::crop(sim_pu_raster, c(0, 0.3, 0, 0.3))
ply <- sim_pu_polygons[c(1:3, 11:13, 20:22), ]
# create boundary matrix using raster data
bm_raster <- boundary_matrix(r)
# create boundary matrix using polygon data
bm_ply <- boundary_matrix(ply)
# plot raster and boundary matrix
plot(r, main = "raster", axes = FALSE)
Matrix::image(bm_raster, main = "boundary matrix")
# plot polygons and boundary matrices
plot(ply[, 1], main = "polygons", axes = FALSE)
Matrix::image(bm_ply, main = "boundary matrix")
# }
```