How to save levelplot() with like cells merged

Question

Say I have a heatmap with contours like so:

set.seed(1)
X <- matrix(runif(100*200),nrow=100)
X <- apply(X,1,sort)
#png("Surface.png",width=800,height=400)
lattice::levelplot(t(X),contour=TRUE)
#dev.off()
getwd()

it looks like this:

There are many ways to make variants of this in R (fields::image.plot(), image(), and so forth. If I save this filled contour plot as a pdf, it's a squeaky-clean vector image that I can rescale and include in a conference poster (using e.g., Inkscape). It turns out in this case that the like-colored areas are actually individual raster cells rather than merged polygons implying a very large number of vertices, which hogs memory and slows down Inkscape after including a couple such surfaces in the poster. The easiest solution would be so save out to the ideal merged-cell format from R, where the contiguous like-colored areas are unified/merged polygon()s or similar, thereby decreasing the number of vertices by an order of magnitude or more.

My question is whether there is some surface function that already does this by default, or a low effort way to emulate this kind of surface output. The high-effort approach would be to dive into R's spatial functions to merge like cells, but I'd rather avoid this. Thanks in advance!

Answer 1

Here is the "high-effort approach" (not really)

set.seed(1)
X <- matrix(runif(100*200),nrow=100)
X <- apply(X,1,sort)
lattice::levelplot(X,contour=TRUE)

library(raster)
r <- raster(X)
z <- cut(r, seq(0, 1, 0.1))
p <- rasterToPolygons(z, dissolve=TRUE)
spplot(p)