Convert a 3D Kernel Density Estimate (package ks) into rgl mesh3d object

Convert a 3D Kernel Density Estimate (package ks) into rgl mesh3d object

# S3 method for kde
as.mesh3d(x, cont = 50, approx.cont = TRUE,
  method = c("tri", "ashape3d"), ashape.alpha = NULL, ...)

Arguments

x	The kde object
cont	The contour level to use
approx.cont	Whether to compute an approximate contour level. (Default FALSE).
method	Which conversion method to use (see details)
ashape.alpha	A length parameter (alpha) passed to alphashape3d::ashape3d (see details).
...	Additional rgl rendering parameters eventually passed to rgl::tmesh3d

Value

An object of class mesh3d

Details

There are two conversion methods. The default uses the output of the misc3d::contour3d function. This looks nice but seems to me to have a lot of redundant triangles. The alternative uses the alphashape3d library which produces a simpler output. But you must provide a length parameter ashape.alpha which is passed to alphashape3d

See also

plot.kde, as.mesh3d, as.mesh3d.ashape3d

Examples

# NOT RUN {
# Assuming you have lh.voxels object from Alex
lh.voxels.m3d=lapply(lh.voxels, function(x) as.mesh3d(attr(x, 'fhat')))
shade3d(lh.voxels.m3d[[1]], col='red', alpha=.3)
# plot several at once
mapply(shade3d, lh.voxels.m3d[c(1,4,10)], col=rainbow(3), alpha=.3)

# alternatively making a smoother/simpler alphashape
# note that the parameter
am1=as.mesh3d(attr(lh.voxels[[1]], 'fhat'), method='alpha', alpha=10)
lh.voxels.m3d2=lapply(, function(x) )
# }