More

Reclassifying raster stack based on condition and other layers using R?

Reclassifying raster stack based on condition and other layers using R?


I want to reclassify rasterr1in the following form:

  • all the values that are larger thanr2to be1
  • all the values that are less thanr3to be0
  • all the rest of the values to be equal tor4

I use overlay to set the values for the first two conditions, but I can't do that for the third one. I am also interested to know how to put all these in one function.

library(raster) r1 <- raster(nrow=5, ncol=5) r1 <- setValues(r1, runif(ncell(r1))) r2 <- setValues(r1, runif(25 ,0.6,0.9)) r3 <- setValues(r1, runif(25 ,0.2,0.4)) r4 <- setValues(r1, runif(25 ,1,2)) x <- overlay(r1, r2, fun=function(x,y){ x[x>y] <- 1; x}) x2 <- overlay(x, r3, fun=function(x,y){ x[x

To put this into a vectorized function you can use ifelse. If you stack your rasters then you do not need to piecemeal the reclassification and can apply a function to the stack.

Prepare data

library(raster) r1 <- raster(nrow=5, ncol=5) r <- stack(setValues(r1, runif(ncell(r1))), setValues(r1, runif(25 ,0.6,0.9)), setValues(r1, runif(25 ,0.2,0.4)), setValues(r1, runif(25 ,1,2)))

Write reclassification function

rc <- function(x1,x2,x3,x4) { ifelse( x1 > x2, 1, ifelse( x1 < x3, 0, x4) ) }

Apply function to raster stack

r.class <- overlay(r, fun=rc)

The bracket extract operators work on rasters in R. So you dont need to use setValues or overlay. I would do this:

r1[r1 > r2] <- 1 r1[r1 < r3] <- 0 r1[!(r1 > r2 | r1 < r3)] <- r4[!(r1 > r2 | r1 < r3)]

This should work if all of your rasters are the same extent and resolution.


Watch the video: CS50 2013 - Week 5, continued