Using patterns in addition/instead of background colors in lattice plots

Question

I am using level plots from the R lattice package. My resulting plots look like the one shown below.

My problem now is that I need to generate a black and white version for printing.

Is there a way to change the colors to grayscale and give the rectangles a background pattern so the the red once are distinguishable from the blue ones? For example, dots or diagonal dashes come to mind.

Thanks!

Answer 1

I found a way to manually draw into the levelplot panel and to draw a diagonal fill pattern over all cells with values greater than 0.5

However, I couldn't manage to draw the same pattern in the color key legend. After hours of reading forums and trying to understand the lattice source code, I couldn't get a clue. Maybe someone else could fix that. Here is what I got:

library(lattice)
library(RColorBrewer)
cols <- colorRampPalette(brewer.pal(8, "RdBu"))

data <- Harman23.cor$cov    

fx <- fy <- c()
for (r in seq(nrow(data)))
  for (c in seq(ncol(data)))
  {
    if (data[r, c] > 0.5)
    {
      fx <- c(fx, r);
      fy <- c(fy, c);
    }
  }

diag_pattern <- function(...)
{
  panel.levelplot(...)
  for (i in seq(length(fx)))
  {
    panel.linejoin(x = c(fx[i],fx[i]+.5), y= c(fy[i]+.5,fy[i]), col="black")
    panel.linejoin(x = c(fx[i]-.5,fx[i]+.5), y= c(fy[i]+.5,fy[i]-.5), col="black")
    panel.linejoin(x = c(fx[i]-.5,fx[i]), y= c(fy[i],fy[i]-.5), col="black")   
  }
}      

p <- levelplot(data, scales=list(x=list(rot=45)), 
               xlab="", ylab="", col.regions=cols, panel=diag_pattern)
print(p)