Plot an interp or tripack Delaunay triangulation

The tri.mesh() functions in the interp and tripack packages compute a Delaunay triangulation of a set of points. These functions display it as a surface.

# S3 method for class 'triSht'
plot3d(x, z, ...)
# S3 method for class 'triSht'
persp3d(x, z, ..., add = FALSE)
# S3 method for class 'triSht'
as.mesh3d(x, z, col = "gray", coords = c("x", "y", "z"), 
    smooth = TRUE, normals = NULL, texcoords = NULL, ...)
# S3 method for class 'tri'
plot3d(x, z, ...)
# S3 method for class 'tri'
persp3d(x, z, ..., add = FALSE)
# S3 method for class 'tri'
as.mesh3d(x, z, col = "gray", coords = c("x", "y", "z"), 
    smooth = TRUE, normals = NULL, texcoords = NULL, ...)

Arguments

x

A "triSht" or "tri" object, produced by the tri.mesh() function in the interp or tripack packages respectively.

z

z coordinate values corresponding to each of the nodes in x.

add

Whether to add surface to existing plot (add = TRUE) or create a new plot (add = FALSE, the default).

col

Colors to apply to each vertex in the triangulation. Will be recycled as needed.

coords

See Details below.

smooth

Whether to average normals at vertices for a smooth appearance.

normals

User-specified normals at each vertex. Requires smooth = FALSE.

texcoords

Texture coordinates at each vertex.

...

See Details below.

Details

These functions construct a mesh3d object corresponding to the triangulation in x. The plot3d and persp3d methods plot it.

The coords parameter allows surfaces to be plotted over any coordinate plane. It should be a permutation of the column names c("x", "y", "z"). The first will be used as the x coordinate, the second as the y coordinate, and the third as the z coordinate.

The ... parameters in plot3d.triSht and plot3d.tri are passed to persp3d; in persp3d.triSht and persp3d.tri they are passed to both as.mesh3d and persp3d.mesh3d; in as.mesh3d.triSht and as.mesh3d.tri they are used as material parameters in a tmesh3d call.

"tri" objects may contain constraints. These appear internally as extra nodes, representing either the inside or outside of boundaries on the region being triangulated. Each of these nodes should also have a z value, but triangles corresponding entirely to constraint nodes will not be drawn. In this way complex, non-convex regions can be triangulated. See the second example below.

Note

If there are duplicate points, the tri.mesh() functions will optionally delete some of them. If you choose this option, the z values must correspond to the nodes after deletion, not before.

Examples

x <- rnorm(200, sd = 5)
y <- rnorm(200, sd = 5)
r <- sqrt(x^2 + y^2)
z <- 10 * sin(r)/r
col <- cm.colors(20)[1 + round(19*(z - min(z))/diff(range(z)))]
save <- NULL
if ((haveinterp <- requireNamespace("interp", quietly = TRUE))) {
  save <- options(rgl.meshColorWarning = FALSE)
  dxy <- interp::tri.mesh(x, y)
  open3d()
  persp3d(dxy, z, col = col, meshColor = "vertices")
}


if (haveinterp) {
  open3d()
  # Do it without smoothing and with a different orientation.
  persp3d(dxy, z, col = col, coords = c("z", "x", "y"), smooth = FALSE)
}


if (requireNamespace("tripack", quietly = TRUE)) {
  if (is.null(save))
    save <- options(rgl.meshColorWarning = FALSE)

  # Leave a circular hole around (3, 0)
  theta <- seq(0, 2*pi, length.out = 30)[-1]
  cx <- 2*cos(theta) + 3
  cy <- 2*sin(theta)
  keep <- (x - 3)^2 + y^2 > 4
  dxy2 <- tripack::tri.mesh(x[keep], y[keep])
  dxy2 <- tripack::add.constraint(dxy2, cx, cy)
  z <- dxy2$x^2 - dxy2$y^2
  col <- terrain.colors(20)[1 + round(19*(z - min(z))/diff(range(z)))]
  open3d()
  persp3d(dxy2, z, col = col)
}
#> Registered S3 methods overwritten by 'tripack':
#>   method                from  
#>   plot.voronoi          interp
#>   plot.voronoi.polygons interp
#>   print.voronoi         interp
#>   print.summary.voronoi interp
#>   summary.voronoi       interp


options(save)