demo/spatstat.R

In spatstat: Spatial Point Pattern Analysis, Model-Fitting, Simulation, Tests

if(dev.cur() <= 1) {
  dd <- getOption("device")
  if(is.character(dd)) dd <- get(dd)
  dd()
}

oldpar <- par(ask = interactive() && dev.interactive(orNone=TRUE))
oldoptions <- options(warn=-1)

fanfare <- function(stuff) {
  plot(c(0,1),c(0,1),type="n",axes=FALSE, xlab="", ylab="")
  text(0.5,0.5, stuff, cex=2.5)
}

par(mar=c(1,1,2,1)+0.1)

fanfare("Spatstat demonstration")

fanfare("I. Types of data")
plot(swedishpines, main="Point pattern")

plot(demopat, cols=c("green", "blue"), main="Multitype point pattern")

plot(longleaf, fg="blue", main="Marked point pattern")

plot(finpines, main="Point pattern with multivariate marks")

a <- psp(runif(20),runif(20),runif(20),runif(20), window=owin())
plot(a, main="Line segment pattern")
marks(a) <- sample(letters[1:4], 20, replace=TRUE)
plot(a, main="Multitype line segment pattern")
marks(a) <- runif(20)
plot(a, main="Marked line segment pattern", style="width")

plot(owin(), main="Rectangular window")
plot(letterR, main="Polygonal window")
plot(as.mask(letterR), main="Binary mask window")

Z <- as.im(function(x,y){ sqrt((x - 1)^2 + (y-1)^2)}, square(2))
plot(Z, main="Pixel image")

X <- runifpoint(42)
plot(dirichlet(X), main="Tessellation")

plot(rpoispp3(100), main="Three-dimensional point pattern")

plot(simplenet, main="Linear network (linnet)")

X <- rpoislpp(20, simplenet)
plot(X,
     main="Point pattern on linear network (lpp)",
     show.window=FALSE)

V <- as.linim(function(x,y,seg,tp){x^2-y^2}, L=simplenet)
plot(V, main="Pixel image on a linear network")

fanfare("II. Graphics")

par(mar=c(5,4,2,3)+0.1)

plot(letterR, col="green", border="red", lwd=2, main="Polygonal window with colour fill")
plot(letterR, hatch=TRUE, spacing=0.15, angle=30, main="Polygonal window with line shading")
plot(letterR, hatch=TRUE, hatchargs=list(texture=8, spacing=0.12),
     main="Polygonal window with texture fill")

plot(amacrine, chars=c(1,16),
     main="plot(X, chars = c(1,16))")
plot(amacrine, cols=c("red","blue"), chars=16,
     main="plot(X, cols=c(\"red\", \"blue\"))")

opa <- par(mfrow=c(1,2))
plot(longleaf, markscale=0.03, main="markscale=0.03")
plot(longleaf, markscale=0.09, main="markscale=0.09")           
par(opa)

plot(longleaf, pch=21, cex=1,
     bg=colourmap(terrain.colors(128), range=c(0,80)),
     main="colourmap for numeric mark values")

Z <- as.im(function(x,y) { r <- sqrt(x^2+y^2); r * exp(-r) },
           owin(c(-5,5),c(-5,5)))
plot(Z, main="pixel image: image plot")
plot(Z, main="pixel image: image plot (heat colours)", col=heat.colors(256))
plot(Z, main="pixel image: logarithmic colour map", 
     log=TRUE, col=rainbow(128, end=5/6))
contour(Z, main="pixel image: contour plot", axes=FALSE)
plot(Z, main="pixel image: image + contour plot")
contour(Z, add=TRUE)
persp(Z, colmap=terrain.colors(128), shade=0.3, phi=30,theta=100,
      main="pixel image: perspective plot")

ct <- colourmap(rainbow(20), breaks=seq(-1,1,length=21))
plot(ct, main="Colour map for real numbers")

ca <- colourmap(rainbow(8), inputs=letters[1:8])
plot(ca, main="Colour map for discrete values")

Z <- as.im(nnfun(runifpoint(8)))
plot(Z, main="colour image for discrete values")
textureplot(Z, main="texture plot for discrete values")

W <- owin(c(1,5),c(0,4.5))
Lout <- scaletointerval(distmap(rebound.owin(letterR, W)))
Lin <- scaletointerval(distmap(complement.owin(letterR, W)))
L <- scaletointerval(eval.im(Lin-Lout))
D <- scaletointerval(density(runifpoint(30, W), adjust=0.3))
X <- scaletointerval(as.im(function(x,y){ x }, W=W))
plot(listof(L=L, D=D, X=X), main="Multiple images")
pairs(L, D, X, main="Multiple images: pairs plot")
persp(L, colin=D,
      theta=-24, phi=35, box=FALSE, apron=TRUE, 
      main="Two images:\nperspective + colours",
      shade=0.4, ltheta=225, lphi=10)
plot(rgbim(D,X,L,maxColorValue=1), valuesAreColours=TRUE,
     main="Three images: RGB display")
plot(hsvim(D,L,X), valuesAreColours=TRUE,
     main="Three images: HSV display")

V <- as.linim(function(x,y,seg,tp){(y/1000)^2-(x/1000)^3}, L=domain(chicago))
plot(V, main="Pixel image on a linear network (colour plot)")
plot(V, style="w", main="Pixel image on a linear network (width plot)")
persp(V, phi=25, theta=-35, expand=0.2, main="Pixel image on a linear network (perspective plot)")

XL <- rpoislpp(20, simplenet)
plot(XL, main="Point pattern on a network (default plot)")
plot(XL,
     shape="crossticks", size=0.04,
     main="Point pattern on a network (crossticks plot)")

plot(chicago, cols=1:7,
     main="Multitype point pattern on a network (symbols and colours)")
plot(dendrite, shape="crossticks", cols=2:4, size=8,
     leg.side="bottom", leg.args=list(lwd=3),
     main="Multitype point pattern on a network (crossticks and colours)")


fanfare("III. Conversion between types")

W <- as.owin(chorley)
plot(W, "window W")

plot(as.mask(W))
plot(as.mask(W, dimyx=1000))

plot(as.im(W, value=3))
plot(as.im(W, value=3, na.replace=0), ribbon=TRUE)

plot(as.im(function(x,y) {x^2 + y}, W=square(1)),
     main="as.im(function(x,y){x^2+y})")

V <- delaunay(runifpoint(12))
plot(V, main="Tessellation V")
plot(as.im(V, dimyx=256), main="as.im(V)")
plot(as.owin(V))

X <- swedishpines
plot(X, "point pattern X")

plot(as.im(X), col=c("white","red"), ribbon=FALSE, xlab="", ylab="")
plot(as.owin(X), add=TRUE)

fanfare("IV. Subsetting and splitting data")

plot(X, "point pattern X")
subset <- 1:20
plot(X[subset], main="subset operation: X[subset]")
subwindow <- owin(poly=list(x=c(0,96,96,40,40),y=c(0,0,100,100,50)))
plot(X[subwindow], main="subset operation: X[subwindow]")

plot(lansing, "Lansing Woods data")
plot(split(lansing),
     main="split operation: split(X)",
     mar.panel=c(0,0,2,0), hsep=1, pch=3)

plot(longleaf, main="Longleaf Pines data")
plot(cut(longleaf, breaks=3),
     main=c("cut operation", "cut(longleaf, breaks=3)"),
     cols=2:4)

Z <- dirichlet(runifpoint(10))
X <- runifpoint(100)

plot(cut(X,Z), main="points cut by tessellation", leg.side="left", cols=1:10)
plot(Z, add=TRUE)

plot(split(X, Z),
     main="points split by tessellation",
     mar.panel=c(0,0,2,2), hsep=1)

W <- square(1)
X <- as.im(function(x,y){sqrt(x^2+y^2)}, W)
Y <- dirichlet(runifpoint(12, W))
plot(split(X,Y), main="image split by tessellation")

fanfare("V. Tessellations")

plot(quadrats(square(1), nx=5, ny=3))
plot(hextess(square(1), 0.07))
plot(quantess(letterR, "x", 7))
plot(polartess(letterR, nangular=6, radii=(0:4)/2, origin=c(2.8, 1.5)),
     do.col=TRUE)
plot(delaunay(cells))
plot(cells, add=TRUE)
plot(dirichlet(cells))
plot(cells, add=TRUE)
plot(rpoislinetess(2.5), do.col=TRUE)

fanfare("VI. Exploratory data analysis")

par(mar=c(3,3,3,2)+0.1)

plot(swedishpines, main="Quadrat counts", pch="+")
tab <- quadratcount(swedishpines, 4)
plot(tab, add=TRUE, lty=2, cex=2, col="blue")

par(mar=c(5,3,3,2)+0.1)

plot(swedishpines, main="", pch="+")
title(main=expression(chi^2 * " test"), cex.main=2)
tes <- quadrat.test(swedishpines, 3)
tes
plot(tes, add=TRUE, col="red", cex=1.5, lty=2, lwd=3)
title(sub=paste("p-value =", signif(tes$p.value,3)), cex.sub=1.4)

par(mar=c(4,4,3,2)+0.1)

tesk <- cdf.test(nztrees, "x")
tesk
plot(tesk)


mur <- lapply(murchison, rescale, s=1000)
mur <- lapply(mur, "unitname<-", value="km")
X <- mur$gold
D <- distfun(mur$faults)
plot(X, main="Murchison gold deposits", cols="blue")
plot(mur$faults, add=TRUE, col="red")
rh <- rhohat(X,D)
plot(rh,
     main="Smoothed rate estimate",
     xlab="Distance to nearest fault (km)",
     legend=FALSE, xlim=c(0, 20))
plot(predict(rh), main="predict(rhohat(X,D))")

Z <- density(cells, 0.07)
plot(Z, main="Kernel smoothed intensity of point pattern")
plot(cells, add=TRUE)

plot(redwood, main="Redwood data")
te <- scan.test(redwood, 0.1, method="poisson")
plot(te, main=c("Scan Statistic for redwood data",
              paste("p-value =", signif(te$p.value,3))))
plot(redwood, add=TRUE)
te

X <- unique(unmark(shapley))
plot(X, "Shapley galaxy concentration", pch=".")
coco <-colourmap(rev(rainbow(128, end=2/3)), range=c(0,1))
pa <- function(i, ...) {
  if(i == 1) list(chars=c(".", "+"), cols=1:2) else
             list(size=0.5, pch=16, col=coco)
}
plot(nnclean(X, k=17), panel.args=pa,
     mar.panel=c(0,1,1,0), nrows=2,
     main="Byers-Raftery nearest neighbour cleaning",
     cex.title=1.2)
Y <- sharpen(X, sigma=0.5, edgecorrect=TRUE)
plot(Y, main="Choi-Hall data sharpening", pch=".")

  owpa <- par(mfrow=c(1,2))
  W <- grow.rectangle(as.rectangle(letterR), 1)
  X <- superimpose(runifpoint(300, letterR),
                   runifpoint(50, W), W=W)
  plot(W, main="clusterset(X, 'm')")
  plot(clusterset(X, 'marks', fast=TRUE), add=TRUE, chars=c("o", "+"), cols=1:2)
  plot(letterR, add=TRUE)
  plot(W, main="clusterset(X, 'd')")
  plot(clusterset(X, 'domain', exact=FALSE), add=TRUE)
  plot(letterR, add=TRUE)
  par(owpa)

D <- density(a, sigma=0.05)
plot(D, main="Kernel smoothed intensity of line segment pattern")
plot(a, add=TRUE)

X <- runifpoint(42)
plot(dirichlet(X))
plot(X, add=TRUE)

plot(delaunay(X))
plot(X, add=TRUE)

parsave <- par(mfrow=c(1,1), mar=0.2+c(0,1,3,1))
plot(listof("Longleaf Pines data"=longleaf,
            "Nearest mark"=nnmark(longleaf),
            "Kernel smoothing of marks"=Smooth(longleaf,10),
            "Inverse distance weighted\nsmoothing of marks"=idw(longleaf)),
     equal.scales=TRUE, halign=TRUE, valign=TRUE,
     main="", mar.panel=0.2+c(0,0,2,2))
par(parsave)

fryplot(cells, main=c("Fry plot","cells data"), pch="+")
miplot(longleaf, main="Morishita Index plot", pch=16, col="blue")

plot(swedishpines, main="Swedish Pines data")
K <- Kest(swedishpines)
plot(K, main="K function for Swedish Pines", legendmath=TRUE)

en <- envelope(swedishpines, fun=Kest, nsim=10, correction="translate")
plot(en, main="Envelopes of K function based on CSR", shade=c("hi", "lo"))

pc <- pcf(swedishpines)
plot(pc, main="Pair correlation function")

plot(swedishpines, main="nearest neighbours")
m <- nnwhich(swedishpines)
b <- swedishpines[m]
arrows(swedishpines$x, swedishpines$y, b$x, b$y,
       angle=12, length=0.1, col="red")

plot(swedishpines %mark% nndist(swedishpines),
     markscale=1, main="Stienen diagram", legend=FALSE, fg="blue")

plot(Gest(swedishpines),
     main=c("Nearest neighbour distance function G", "Gest(swedishpines)"),
     legendmath=TRUE)

Z <- distmap(swedishpines, dimyx=512)
plot(swedishpines$window, main="Distance map")
plot(Z, add=TRUE)
points(swedishpines)

plot(Fest(swedishpines),
     main=c("Empty space function F", "Fest(swedishpines)"),
     legendmath=TRUE)

W <- rebound.owin(letterR, square(5))
plot(distmap(W), main="Distance map")
plot(W, add=TRUE)

a <- psp(runif(20),runif(20),runif(20),runif(20), window=owin())
contour(distmap(a), main="Distance map")
plot(a, add=TRUE,col="red")

plot(Jest(swedishpines), main=c("J-function", "J(r)=(1-G(r))/(1-F(r))"))

X <- swedishpines
X <- X[sample(1:npoints(X))]
Z <- nnfun(X)
plot(as.owin(X), main="Nearest neighbour map")
plot(Z, add=TRUE)
points(X)

plot(allstats(swedishpines))

Fig4b <- residualspaper$Fig4b

plot(Fig4b, main="Inhomogeneous point pattern")
plot(Kinhom(Fig4b), main="Inhomogeneous K-function")
plot(pcfinhom(Fig4b, stoyan=0.1), main="Inhomogeneous pair correlation")
plot(Ginhom(Fig4b, sigma=0.06), main="Inhomogeneous G-function")
plot(Jinhom(Fig4b, sigma=0.06), main="Inhomogeneous J-function")

X <- unmark(bronzefilter)
plot(X, "Bronze filter data")
lam <- predict(ppm(X ~x))
plot(Kscaled(X, lam), xlim=c(0, 1.5), main="Locally-scaled K function")

plot(urkiola)
plot(split(urkiola), cex=0.5)
plot(density(split(urkiola)))
contour(density(split(urkiola)), panel.begin=as.owin(urkiola))
plot(relrisk(urkiola), main="Relative risk (cross-validated)")

plot(bramblecanes)
br <- rescale(bramblecanes)
plot(alltypes(br, "K"), mar.panel=c(4,5,2,2)+0.1)

ama <- rescale(amacrine)
plot(alltypes(ama, Lcross, envelope=TRUE, nsim=9), . - r ~ r, ylim=c(-25, 5))

ponderosa.extra$plotit(main="Ponderosa Pines")

L <- localL(ponderosa)
pL <- plot(L, lty=1, col=1, legend=FALSE,
           main=c("neighbourhood density functions",
             "for Ponderosa Pines"), cex.main=0.8)

parsave <- par(mfrow=c(1,2))
ponderosa.extra$plotit()
par(pty="s")
plot(L, iso007 ~ r, main="point B")

par(mar=0.2+c(1,1,3,1))
ponderosa.extra$plotit()
L12 <- localL(ponderosa, rvalue=12)
P12 <- ponderosa %mark% L12
Z12 <- Smooth(P12, sigma=5, dimyx=128)
plot(Z12, col=topo.colors(128),
     main=c("smoothed", "neighbourhood density"),
     cex.main=0.8)
contour(Z12, add=TRUE)
points(ponderosa, pch=16, cex=0.5)

plot(amacrine, main="Amacrine cells data", cex.main=0.8)
par(pty="s")
mkc <- markcorr(amacrine, 
                correction="translate", method="density",
                kernel="epanechnikov")
plot(mkc, main="Mark correlation function", legend=FALSE, cex.main=0.8)
par(parsave)

par(mar=0.2+c(4,4,3,1))
plot(alltypes(amacrine, markconnect), 
     title="Mark connection functions for amacrine cells")

parsave <- par(mfrow=c(1,2))

parspruce2 <- par(mar=0.2+c(0,2,2,0))
plot(spruces, cex.main=0.8, markscale=10)
par(pty="s", mar=0.2+c(2,3,2,0))
plot(markcorr(spruces), main="Mark correlation", legendpos="bottomright")

par(parspruce2)
plot(spruces, cex.main=0.8, markscale=10)
par(pty="s", mar=0.2+c(2,3,2,0))
plot(markvario(spruces), main="Mark variogram", legendpos="topright")
par(parsave)
par(mar=0.2+c(4,4,3,2))

plot(listof("Emark(spruces)"=Emark(spruces),
            "Vmark(spruces)"=Vmark(spruces)),
     main="Independence diagnostics", ylim.covers=0,
     legendpos="bottom")

par3 <- par(mfrow=c(1,2))
X <- rpoispp3(100)
plot(X, main="3D point pattern X")
plot(K3est(X), main="K-function in 3D")
plot(X, main="3D point pattern X")
plot(G3est(X), main="G-function in 3D", legendpos="bottomright")
par(par3)

par(mfrow=c(1,3))
X <- unmark(chicago)
plot(X, col="green", cols="red", pch=16,
     main="Chicago Street Crimes", cex.main=0.75,
     show.window=FALSE)
plot(density(X, 100, distance="e"), main="Kernel density estimate (Euclidean)")
plot(density(X, 100, distance="p"), main="Kernel density estimate (shortest path)")

plot(X, col="green", cols="red", pch=16,
     main="Chicago Street Crimes", cex.main=0.75,
     show.window=FALSE)
plot(linearK(X, correction="none"), main="Network K-function", cex.main=0.75)
plot(linearK(X, correction="Ang"), main="Corrected K-function", cex.main=0.75)

par(mfrow=c(1,1))

fanfare("VII. Model-fitting")

parsave <- par(mar=0.2+c(1,1,3,3))
plot(japanesepines)
fit <- ppm(japanesepines ~1)
print(fit)
fit <- ppm(japanesepines ~polynom(x,y,2))
print(fit)
plot(fit, how="image", se=FALSE, main=c("Inhomogeneous Poisson model",
                               "fit by maximum likelihood",
                               "Fitted intensity"))
plot(fit, how="image", trend=FALSE,
     main=c("Standard error", "of fitted intensity"))

plot(leverage(fit))
plot(influence(fit))

plot(mur$gold, main="Murchison gold deposits", cols="blue")
plot(mur$faults, add=TRUE, col="red")
fit <- ppm(mur$gold ~D, covariates=list(D=distfun(mur$faults)))
par(mar=0.2+c(4,4,4,2))
plot(parres(fit, "D"),
     main="Partial residuals from loglinear Poisson model",
     xlab="Distance to nearest fault (km)",
     ylab="log intensity of gold", legend=FALSE)
legend("bottomleft", legend=c("partial residual", "loglinear fit"),
       col=c(1,4), lty=c(1,4))

par(mar=rep(0.2, 4), mfrow=c(1,1))
fitT <- kppm(redwood ~1, clusters="Thomas")
simT <- simulate(fitT)[[1]]
plot(listof(redwood, simT),
     main.panel=c("Redwood", "simulation from\nfitted Thomas model"),
     main="", mar.panel=0.2, equal.scales=TRUE)

mop <- par(mfrow=c(1,2), pty="s", mar=rep(4.4, 4))
plot(fitT, xname=c("Thomas model", "minimum contrast fit"), pause=FALSE)
par(mop)

oop <- par(pty="s", mar=0.2+c(4,4,4,2))
os <- objsurf(fitT)
plot(os, main="Minimum contrast objective function", col=terrain.colors(128))
contour(os, add=TRUE)
par(oop)

parra <- par(mfrow=c(1,2), mar=0.2+c(3,3,4,2))
plot(swedishpines)
fit <- ppm(swedishpines ~1, Strauss(r=7))
print(fit)
plot(fit, how="image", main=c("Strauss model",
                               "fit by maximum pseudolikelihood",
                               "Conditional intensity plot"))
# fitted interaction
plot(swedishpines)
fit <- ppm(swedishpines ~1, PairPiece(c(3,5,7,9,11,13)))
plot(fitin(fit), legend=FALSE,
     main=c("Pairwise interaction model",
            "fit by maximum pseudolikelihood"))

# simulation
par(mfrow=c(1,1), mar=0.5+c(0,0,2,0))
Xsim <- rmh(model=fit,
            start=list(n.start=80),
            control=list(nrep=100))
plot(listof(swedishpines, Xsim),
     main="",
     main.panel=c("Swedish Pines",
       "Simulation from\nfitted Strauss model"),
     mar.panel=c(0,0,3,0),hsep=1,equal.scales=TRUE)

# model compensator
par(parra)
par(mar=0.2+c(4,4,3,1))
plot(swedishpines)
fit <- ppm(swedishpines ~1, Strauss(r=7))
plot(Kcom(fit), cbind(iso, icom, pois) ~ r,
     legend=FALSE, main="model compensators")
legend("topleft", legend=c("empirical K function",
                    "Strauss model compensator of K",
                    "Poisson theoretical K"), lty=1:3, col=1:3, inset=0.05)

par(parsave)

# Multitype data
dpat <- rescale(demopat, 8)
unitname(dpat) <- c("mile", "miles")
dpat

plot(dpat, cols=c("red", "blue"))
fit <- ppm(dpat ~marks + polynom(x,y,2), Poisson())
plot(fit, trend=TRUE, se=TRUE)

# Linear network data
plot(spiders)
fit <- lppm(spiders ~ polynom(x,y,2))
anova(fit, test="Chi")
(fit <- step(fit))
lam <- predict(fit)
plot(lam, main="Point process model on linear network", ribscale=1000)
plot(spiders, add=TRUE, pch=16, show.network=FALSE)

fanfare("VIII. Simulation")

plot(letterR, main="Poisson random points")
lambda <- 10/area.owin(letterR)
points(rpoispp(lambda, win=letterR))
points(rpoispp(9 * lambda, win=letterR))
points(rpoispp(90 * lambda, win=letterR))
plot(rpoispp(100))
plot(rpoispp(function(x,y){1000 * exp(-3*x)}, 1000),
     main="rpoispp(function)")

plot(rMaternII(200, 0.05))
plot(rSSI(0.05, 200))
plot(rThomas(10, 0.2, 5))
plot(rMatClust(10, 0.05, 4))
plot(rCauchy(30, 0.01, 5))
plot(rVarGamma(30, 2, 0.02, 5))
plot(rGaussPoisson(30, 0.05, 0.5))

if(require(RandomFields) && RandomFieldsSafe()) {
  X <- rLGCP("exp", 4, var=0.2, scale=0.1)
  plot(attr(X, "Lambda"), main="log-Gaussian Cox process")
  plot(X, add=TRUE, pch=16)
}

plot(rStrauss(200, 0.3, 0.07))
plot(rDiggleGratton(200,0.03,0.08))
plot(rDGS(300, 0.05))

plot(redwood, main="random thinning - rthin()")
points(rthin(redwood, 0.5), col="green", cex=1.4)

plot(rcell(nx=15))

plot(rsyst(nx=5))
abline(h=(1:4)/5, lty=2)
abline(v=(1:4)/5, lty=2)

plot(rstrat(nx=5))
abline(h=(1:4)/5, lty=2)
abline(v=(1:4)/5, lty=2)

X <- rsyst(nx=10)
plot(rjitter(X, 0.02))

Xg <- rmh(list(cif="geyer", par=list(beta=1.25, gamma=1.6, r=0.2, sat=4.5),
               w=c(0,10,0,10)),
          control=list(nrep=1e4), start=list(n.start=200))
plot(Xg, main=paste("Geyer saturation process\n",
                    "rmh() with cif=\"geyer\""))

L <- as.psp(matrix(runif(20), 5, 4), window=square(1))
plot(L, main="runifpointOnLines(30, L)")
plot(runifpointOnLines(30, L), add=TRUE, pch="+")

plot(L, main="rpoisppOnLines(3, L)")
plot(rpoisppOnLines(3, L), add=TRUE, pch="+")

plot(runiflpp(20, simplenet))
plot(rpoislpp(5, simplenet))

plot(rpoisline(10))

plot(rlinegrid(30, 0.1))

spatstat.options(npixel=256)
X <- dirichlet(runifpoint(30))
plot(rMosaicSet(X, 0.4), col="green", border=NA)
plot(X, add=TRUE)
plot(rMosaicField(X, runif))
plot(rMosaicSet(rpoislinetess(3), 0.5), col="green", border=NA, main="Switzer's random set")
spatstat.options(npixel=100)

plot(Halton(512, c(2,3)), main="quasirandom pattern")
plot(Halton(16384, c(2,3)), main="quasirandom pattern", pch=".")

fanfare("IX. Geometry")

A <- letterR

B <- shift(letterR, c(0.2,0.1))
plot(boundingbox(A,B), main="shift", type="n")
plot(A, add=TRUE)
plot(B, add=TRUE, border="red")

B <- rotate(letterR, 0.2)
plot(boundingbox(A,B), main="rotate", type="n")
plot(A, add=TRUE)
plot(B, add=TRUE, border="red")

mat <- matrix(c(1.1, 0, 0.3, 1), 2, 2)
B <- affine(letterR, mat=mat, vec=c(0.2,-0.1))
plot(boundingbox(A,B), main="affine", type="n")
plot(A, add=TRUE)
plot(B, add=TRUE, border="red")

par1x2 <- par(mfrow=c(1,2))
L <- rpoisline(10, owin(c(1.5,4.5),c(0.2,3.6)))
plot(L, main="Line segment pattern")
plot(L$window, main="L[window]", type="n")
plot(L[letterR], add=TRUE)
plot(letterR, add=TRUE, border="red")
par(par1x2)

a <- psp(runif(20),runif(20),runif(20),runif(20), window=owin())
plot(a, main="Self-crossing points")
plot(selfcrossing.psp(a), add=TRUE, col="red")

a <- as.psp(matrix(runif(20), 5, 4), window=square(1))
b <- rstrat(square(1), 5)
plot(a, lwd=3, col="green", main="project points to segments")
plot(b, add=TRUE, col="red", pch=16)
v <- project2segment(b, a)
Xproj <- v$Xproj
plot(Xproj, add=TRUE, pch=16)
arrows(b$x, b$y, Xproj$x, Xproj$y, angle=10, length=0.15, col="red")

plot(a, main="pointsOnLines(L)")
plot(pointsOnLines(a, np=100), add=TRUE, pch="+")

parry <- par(mfrow=c(1,3), mar=0.3+c(1,1,3,1))
X <- tess(xgrid=seq(2, 4, length=10), ygrid=seq(0, 3.5, length=8))
plot(X, cex.main=0.75)
plot(letterR, cex.main=0.75)
plot(intersect.tess(X, letterR), cex.main=0.75)

X <- dirichlet(runifpoint(10))
plot(X)
L <- infline(0.3,0.5)
plot(owin(), main="L", cex.main=0.75)
plot(L, col="red", lwd=2, cex.main=0.75)
plot(chop.tess(X,L), cex.main=0.75)
par(parry)

W <- chorley$window
plot(W, main="simplify.owin")
WS <- simplify.owin(W, 2)
plot(WS, add=TRUE, border="green")

nopa <- par(mfrow=c(2,2))
Rbox <- grow.rectangle(as.rectangle(letterR), 0.3)

v <- erosion.owin(letterR, 0.25)
plot(Rbox, type="n", main="erode.owin", cex.main=0.75)
plot(letterR, add=TRUE, col="red", cex.main=0.75)
plot(v, add=TRUE, col="blue")

v <- dilation.owin(letterR, 0.25)
plot(Rbox, type="n", main="dilate.owin", cex.main=0.75)
plot(v, add=TRUE, col="blue")
plot(letterR, add=TRUE, col="red")

v <- closing.owin(letterR, 0.3)
plot(Rbox, type="n", main="closing.owin", cex.main=0.75)
plot(v, add=TRUE, col="blue")
plot(letterR, add=TRUE, col="red")

v <- opening.owin(letterR, 0.3)
plot(Rbox, type="n", main="opening.owin", cex.main=0.75)
plot(letterR, add=TRUE, col="red")
plot(v, add=TRUE, col="blue")
par(nopa)

laslett(heather$fine, main="Laslett's Transform")

fanfare("X. Operations on pixel images")

Z <- distmap(swedishpines, dimyx=512)
plot(Z, main="An image Z")
plot(levelset(Z, 4))
plot(cut(Z, 5))
plot(eval.im(sqrt(Z) - 3))
plot(solutionset(abs(Z - 6) <= 1))
nopa <- par(mfrow=c(1,2))
plot(Z)
segments(0,0,96,100,lwd=2)
plot(transect.im(Z))
par(nopa)

d <- distmap(cells, dimyx=256)
W <- levelset(d, 0.06)
nopa <- par(mfrow=c(1,2))
plot(W)
plot(connected(W))
par(nopa)

Z <- as.im(function(x,y) { 4 * x^2 + 3 * y }, letterR)
plot(Z)
plot(letterR, add=TRUE)

plot(blur(Z, 0.3, bleed=TRUE))
plot(letterR, add=TRUE)

plot(blur(Z, 0.3, bleed=FALSE))
plot(letterR, add=TRUE)
          
plot(blur(Z, 0.3, bleed=FALSE))
plot(letterR, add=TRUE)
          
fanfare("XI. Programming tools")

showoffK <- function(Y, current, ..., fullpicture,rad) { 
	plot(fullpicture,
             main=c("Animation using `applynbd'", "explaining the K function"))
	points(Y, cex=2)
        ux <- current[["x"]]
        uy <- current[["y"]]
	points(ux,uy,pch="+",cex=3)
	theta <- seq(0,2*pi,length=100)
	polygon(ux + rad * cos(theta), uy + rad*sin(theta))
	text(ux+rad/3,uy+rad/2,Y$n,cex=3)
        if(runif(1) < 0.2) Sys.sleep(runif(1, max=0.4))
	return(npoints(Y))
}
par(ask=FALSE)
applynbd(redwood, R=0.2, showoffK, fullpicture=redwood, rad=0.2, exclude=TRUE)

par(oldpar)
options(oldoptions)