poly.network {doolkit}

Identify polygon networks

Description

From a selected variable y, identifies patches of adjacent polygons that share a given range of y values. These patches are called ’polygon networks’.

Usage

poly.network(
  mesh,
  y,
  lwr.limit = stats::quantile(y, 0.75),
  upr.limit = stats::quantile(y, 1),
  min.size = 3
)

Arguments

mesh

object of class mesh3d

y

a vector of indices to be used to select polygons

lwr.limit

the lower range of values to be selected from y

upr.limit

the upper range of values to be selected from y

min.size

the minimum amount of polygons defining a cluster. Default is set to 3.

Value

An object of class "polygon.network" composed of the face index and the membership of each triangle answering the set conditions. The function makes patches of contiguous triangles, and each patch is indexed with a unique number corresponding to its membership.

Examples

#Isolate cusps using elevation:
mythreshold <- quantile(elev(dkmodel$cusp), 0.65)
cusps <- poly.network(dkmodel$cusp, elev(dkmodel$cusp), lwr.limit = mythreshold,
min.size = 100)
myvector <- rep(0, Rvcg::nfaces(dkmodel$cusp))
myvector[cusps@faces] <- cusps@membership[]
myvector <- as.factor(myvector)
ncusps <- length(levels(myvector)) - 1
levels(myvector) <- c(0:ncusps + 1)
dkmap(dkmodel$cusp, as.numeric(myvector), col = cbPalette <- c("#000000", "#E69F00",
"#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7"),
col.levels = ncusps + 1, legend.lab = "Elevation (mm)")

#Any other variables could be used to define the clusters
#Mean curvature:
crests <- poly.network(dkmodel$complex, Rvcg::vcgCurve(dkmodel$complex)$meanitmax,
lwr.limit = quantile(Rvcg::vcgCurve(dkmodel$complex)$meanitmax, 0.8), min.size = 10)
doolkit::dkmap(mesh = dkmodel$complex, y = doolkit::arc(dkmodel$complex,
range = c(-20, 20)), col = "arc", col.levels = 256, min.range = -20,
max.range = 20, orient = "occlusal", legend.lab = "ARC",
alpha.thresh = quantile(doolkit::arc(dkmodel$complex), 0.8), alpha = 0.3,
alpha.above = FALSE)

#Orientation and surface of patches:
patch_orient <- data.frame(bin = NULL, patch = NULL, size = NULL, surface = NULL)
for (i in 1:8) {
  Cluster <- poly.network(dkmodel$complex, orient(dkmodel$complex),
  lwr.limit = 45 * (i - 1), upr.limit = 45 * i)
  Patches <- levels(as.factor(Cluster@membership))
  Bins <- rep(paste(45 * (i - 1), "-", 45 * i), length(Patches))
  Areas <- rep(0, length(Patches))
  for (j in 1:length(Patches)) {
    test <- Cluster@faces[Cluster@membership == Patches[j]]
    Areas[j] <- round(sum(Rvcg::vcgArea(dkmodel$complex,
    perface = TRUE)$pertriangle[test]), 3)
  }
  patch_orient <- data.frame(rbind(patch_orient,
  cbind.data.frame(Bins, Patches, Areas)))
}
#Since patches made of 3 or less polygons are discarded,
#sum of patch areas < total surface area:
sum(patch_orient$Areas)
Rvcg::vcgArea(dkmodel$complex)