Compute cell (pixel) level "local" statistics across layers, and return a single "summary" layer (except for range which returns two layers).
The following summary methods are available for SpatRaster: any, all, max, min, mean, median, prod, range, stdev, sum, which.min, which.max
See app to compute statistics that are not included here or any other custom function.
See modal to compute the mode.
Because generic functions are used, the method applied is chosen based on the first argument: "x". This means that if r is a SpatRaster, mean(r, 5) will work, but mean(5, r) will not work.
The mean method has an argument "trim" that is ignored.
The stdev method returns the population standard deviation, computed as:
f <- function(x) sqrt(sum((x-mean(x))^2) / length(x))
This is different than the sample standard deviation returned by sd (which uses n-1 as denominator). Function f above is equivalent to function g below
g <- function(x) sqrt(sum((x-mean(x))^2) / length(x))