Fast masking of raster values

This is a fast implementation of raster::mask(). It masks values in a Raster object according to values in another Raster or polygon layer. Note that this function might sometimes be slower than raster::mask() when the mask is a Raster. Largest speed gains occur for polygon masks.

fast_mask(ras = NULL, mask = NULL, inverse = FALSE, updatevalue = NA)

Arguments

ras	Raster* object.
mask	Raster* object or POLYGON/MULTIPOLYGON sf::sf() object.
inverse	Logical. If TRUE, areas in mask that are not NA are masked
updatevalue	Convert all masked cells in ras to this value (default is NA)

Value

Raster* object

See also

raster::mask()

Examples

library(raster)

size <- 100
ras <- raster(ncol = size, nrow = size, vals = rep(1, size*size))
msk <- raster(ncol = size, nrow = size, vals = runif(size*size))
msk[msk < 0.5] <- NA

# Using raster as mask

system.time(ras.masked <- raster::mask(ras, msk))
#>    user  system elapsed 
#>    0.01    0.00    0.02 
system.time(ras.masked <- fast_mask(ras, msk))
#>    user  system elapsed 
#>    0.02    0.00    0.02 

plot(ras)
plot(msk)
plot(ras.masked)


# Using polygon layer as mask

msk.sp <- as(msk, 'SpatialPolygonsDataFrame')
msk.sf <- sf::st_as_sf(msk.sp)
system.time(ras.masked <- raster::mask(ras, msk.sp))
#>    user  system elapsed 
#>   12.90    0.06   13.21 
system.time(ras.masked <- raster::mask(ras, msk.sf))
#>    user  system elapsed 
#>   14.01    0.03   14.29 
system.time(ras.masked <- fast_mask(ras, msk.sf))
#> although coordinates are longitude/latitude, st_intersection assumes that they are planar
#>    user  system elapsed 
#>    0.70    0.00    0.72